Development of Lc/Ms Techniques for Plant and Drug Metabolism Studies

A 574 OULU 2011 A 574

UNIVERSITY OF OULU P.O.B. 7500 FI-90014 UNIVERSITY OF OULU FINLAND ACTA UNIVERSITATISUNIVERSITATIS OULUENSISOULUENSIS ACTA UNIVERSITATIS OULUENSIS ACTAACTA

SERIES EDITORS SCIENTIAESCIENTIAEA A RERUMRERUM Aleksanteri Petsalo NATURALIUMNATURALIUM

ASCIENTIAE RERUM NATURALIUM Aleksanteri Petsalo Senior Assistant Jorma Arhippainen DEVELOPMENT OF BHUMANIORA LC/MS TECHNIQUES Lecturer Santeri Palviainen CTECHNICA FOR PLANT AND DRUG Professor Hannu Heusala METABOLISM STUDIES DMEDICA Professor Olli Vuolteenaho ESCIENTIAE RERUM SOCIALIUM Senior Researcher Eila Estola FSCRIPTA ACADEMICA Director Sinikka Eskelinen GOECONOMICA Professor Jari Juga

EDITOR IN CHIEF Professor Olli Vuolteenaho PUBLICATIONS EDITOR

Publications Editor Kirsti Nurkkala UNIVERSITY OF OULU, FACULTY OF SCIENCE, DEPARTMENT OF CHEMISTRY, ISBN 978-951-42-9440-2 (Paperback) FACULTY OF MEDICINE, ISBN 978-951-42-9441-9 (PDF) INSTITUTE OF BIOMEDICINE, ISSN 0355-3191 (Print) DEPARTMENT OF PHARMACOLOGY AND TOXICOLOGY ISSN 1796-220X (Online)

ACTA UNIVERSITATIS OULUENSIS A Scientiae Rerum Naturalium 574

ALEKSANTERI PETSALO

DEVELOPMENT OF LC/MS TECHNIQUES FOR PLANT AND DRUG METABOLISM STUDIES

Academic dissertation to be presented with the assent of the Faculty of Science of the University of Oulu for public defence in Auditorium F100, Futura, Joensuu Campus, on 4 June 2011, at 2 p.m.

Supervised by Docent Ari Tolonen Docent Miia Turpeinen

Reviewed by Professor Janne Jänis Docent Tiia Kuuranne

ISBN 978-951-42-9440-2 (Paperback) ISBN 978-951-42-9441-9 (PDF) http://herkules.oulu.fi/isbn9789514294419/ ISSN 0355-3191 (Printed) ISSN 1796-220X (Online) http://herkules.oulu.fi/issn03553191/

Cover Design Raimo Ahonen

JUVENES PRINT TAMPERE 2011 Petsalo, Aleksanteri, Development of LC/MS techniques for plant and drug metabolism studies. University of Oulu, Faculty of Science, Department of Chemistry, P.O. Box 3000, FI-90014 University of Oulu, Finland, Faculty of Medicine, Institute of Biomedicine, Department of Pharmacology and Toxicology, P.O. Box 5000, FI-90014 University of Oulu, Finland Acta Univ. Oul. A 574, 2011 Oulu, Finland

Abstract Liquid chromatography (LC) combined with mass spectrometry (MS) is a powerful tool for qualitative and quantitative analytics of organic molecules from various matrices, and the use of this hyphenated technique is very common in bioanalytical laboratories. In this study, LC/MS methods and the required sample preparation applications were developed for plant flavonoid and drug metabolism studies. The main focus was in developing methods to be used during cytochrome P450 (CYP) -specific drug interaction studies. Traditional high performance liquid chromatography (HPLC) and new, more efficient and faster ultra-performance liquid chromatography (UPLC) were utilized together with time-of-flight (TOF) and triple quadrupole (QqQ) mass spectrometry. In the flavonoid study, collision-induced radical cleavage of flavonoid glycosides was tested and observed to be a suitable tool for the structure elucidation of the 15 flavonol glycosides extracted from the medicinal plant Rhodiola rosea. Ten of these glycosides were previously unreported in the plant. Several unreported in vivo bupropion metabolites were identified from human urine when developing the method for the new and more extensive in vitro and in vivo N-in-one interaction cocktail assays. The qualified analysis methods developed here enable faster analysis for the N- in-one cocktail assays, in turn enabling a more efficient screening of drugs that affect CYP- enzyme activities. In the case of the human in vitro cocktail assay, fourteen compounds were analyzed using a single LC/MS/MS run. The method has proven to be very reliable and has been used in several interaction studies utilizing different sample matrices. The in vivo cocktail assay that was developed enables totally non-invasive sample collection from the patients, the urine sample being sufficient for the UPLC/MS/MS analysis of all target compounds. The last part of the study consisted of developing a specific and very sensitive UPLC/MS/MS method for the analysis of one of the in vivo cocktail analytes, the antidiabetic drug repaglinide, from human placenta perfusates.

Keywords: cytochrome P450 enzyme system, drug metabolism, flavonoids, liquid chromatography, mass spectrometry, urine analysis

Petsalo, Aleksanteri, LC/MS-menetelmien kehittäminen kasvi- ja lääkeaineiden metabolian tutkimukseen. Oulun yliopisto, Luonnontieteellinen tiedekunta, Kemian laitos, PL 3000, 90014 Oulun yliopisto, Lääketieteellinen tiedekunta, Biolääketieteen laitos, Farmakologia ja toksikologia, PL 5000, 90014 Oulun yliopisto Acta Univ. Oul. A 574, 2011 Oulu

Tiivistelmä Nestekromatografia (LC) yhdistettynä massaspektrometriaan (MS) on tehokas työväline kvalita- tiivisessa ja kvantitatiivisessa analytiikassa, ja tätä tekniikkaa käytetään erityisesti bioalan labo- ratorioissa. Tässä väitöskirjatyössä kehitettiin ja sovellettiin LC/MS- ja näytteenkäsittelymene- telmiä kasvien flavonoidimetabolian ja lääkeaineiden metaboliatuotteiden tutkimukseen keskit- tyen erityisesti sytokromi P450 (CYP) -entsyymispesifisten lääkeaineiden interaktiotutkimuk- siin tarvittaviin menetelmiin. Työssä hyödynnettiin perinteistä korkean erotuskyvyn nestekroma- tografiaa (HPLC) ja uutta, suorituskyvyltään vielä tehokkaampaa ja nopeampaa nestekromato- grafiaa (UPLC) yhdessä lentoaika- (TOF) ja kolmoiskvadrupolimassaspektrometrian (QqQ) kanssa. Tutkimustyön flavonoidimetaboliaan keskittyneessä osuudessa havaittiin törmäyksen aiheuttaman (CID) radikaalipilkkoutumisen soveltuvan lääkinnällisenä kasvina käytetystä ruusujuuresta (Rhodiola rosea) uutettujen viidentoista flavonoliglykosidin rakennemääritykseen. Kymmentä näistä löydetyistä glykosideista ei oltu aiemmin raportoitu ruusujuuresta. Tutkimus- työn keskeisimpänä tavoitteena kehitettiin kvalifioidut LC/MS -analyysimenetelmät käytettäväk- si aikaisempaa kattavampien in vitro ja in vivo -olosuhteiden N-in-one -tyyppisten CYP-entsyy- mi-interaktiotutkimusten analyyttisenä työkaluna. Näitä analyysimenetelmiä kehitettäessä löy- dettiin ja tunnistettiin ihmisen virtsasta aiemmin raportoimattomia metaboliitteja CYP2B6 -entsyymin malliaineena käytetyn bupropionin annostelun jälkeen. Kyseisten kehitettyjen analyysi- menetelmien avulla CYP-entsyymien toimintaan vaikuttavien lääkeaineiden tutkiminen on aiempaa nopeampaa ja antaa yhdellä kertaa samasta tutkimuksesta entistä laaja-alaisempaa tie- toa. In vitro -tutkimusta varten kehitetty LC/MS/MS -analyysimenetelmä on osoittautunut erit- täin käyttökelpoiseksi lukuisissa interaktiotutkimuksissa, ja in vivo -tutkimusta varten kehitetty UPLC/MS/MS -analyysimenetelmä mahdollistaa täysin ei-invasiivisen näytteenoton potilaista. Tutkimustyön viimeisessä vaiheessa kehitettiin erittäin herkkä ja spesifinen UPLC/MS/MS -ana- lyysimenetelmä CYP2C8-entsyymin toiminnan malliaineena käytetyn repaglinidin analysoimi- seksi koejärjestelystä, jossa tutkitaan yhdisteiden kulkeutumista raskausaikana äidin ja sikiön verenkierron välillä istukan kautta.

Asiasanat: aineenvaihdunta, analyysimenetelmät, flavonoidit, lääkeaineet, massaspektrometria, nestekromatografia, sytokromit

To my loved ones 8 Acknowledgements

This work was carried out during the years 2005–2011. Most of the laboratory work was performed at Novamass facilities for which I am very grateful to Jouko Uusitalo (the founder of Novamass) and my supervisor Docent Ari Tolonen for giving me the opportunity to do my PhD studies there. I wish to express my warmest thanks to my supervisors, Docents Ari Tolonen and Miia Turpeinen, for introducing me to the ﬁelds of LC/MS and drug metabolism. Your guidance, knowledge and understanding have been a great support during these years. I am grateful to Professor Janne Jänis and Docent Tiia Kuuranne for their careful review of this thesis. Your constructive criticism and comments were valuable. I also want to thank Dr. Roy Goldblatt for revising the language of this thesis. I would like to thank all my co-authors, and especially professor Olavi Pelkonen for his whole-hearted contribution to our studies. I warmly thank all my colleagues and co-workers at Novamass and at the Structural Elucidation Chemistry Division of the Department of Chemistry for their collaboration. Particularly the laboratory staff is gratefully acknowledged. Finally, I would like to thank all my loved ones. I am deeply grateful to my parents for their never-ending love and support. I thank my brother for being nothing but himself. You have helped me in so many ways. I thank my sister and her family for the fun times we have had. I also want to thank PK for the everlasting support and believe in me. I sincerely thank my godparents and relatives for their ﬁnancial support during these years. With your support I have been able to do things I like. My dear friends, you all are most warmly acknowledged. You make my journey worth of living and I can never express how much you all mean to me. And last but not least, I would like to thank my own coach for her love and understanding. Everything would be so different without you. Funding provided by the Orion-Farmos Research Foundation and Tauno Tönning Foundation is gratefully acknowledged.

Joensuu, April 2011

Aleksanteri Petsalo

9 10 Abbreviations

ADMET absorption, distribution, metabolism, excretion, toxicity ANR anthocyanidin reductase ANS anthocyanidin synthase APCI atmospheric pressure chemical ionization API atmospheric pressure ionization CHI chalcone isomerase CHS chalcone synthase CID collision induced dissociation CRM charge residue model CYP cytochrome P450 DFR dihydroflavonol 4-reductase ESI electrospray ionization F3H flavanone 3-hydroxylase FNS flavone synthase GSH glutathione HPLC high-performance liquid chromatography IEM ion evaporation method IFS isoflavone synthase IS internal standard LC liquid chromatography LLE liquid-liquid extraction LoD limit of detection MDF mass defect filter MRM multiple reaction monitoring MS mass spectrometry MS/MS tandem mass spectrometry m/z mass-to-charge ratio NADPH nicotinamide adenine dinucleotide phosphate NAT N-acetyltransferase NMR nuclear magnetic resonance PAH polycyclic aromatic hydrocarbon

11 PAPS 3’-phosphoadenosine-5’-phosphosulfate PBS phosphate buffered saline PDA photodiode array PP protein precipitation QqQ triple quadrupole RP reversed phase SAA single analyte assay SAM S-adenosylmethionine SFE supercritical ﬂuid extraction SPE solid phase extraction SRM selective reaction monitoring TOF time-of-ﬂight UDP uridine diphosphate UDPGA uridine-5’-diphosphoglucuronic acid UGT UDP-glucuronosyltransferase UPLC ultra-performance liquid chromatography

12 List of original articles

This thesis is based on the following publications, which are referred to in the text by their Roman numerals [I–V]:

I Petsalo A, Jalonen J & Tolonen A (2006) Identification of flavonoids of Rhodiola rosea by liquid chromatography-tandem mass spectrometry. Journal of Chromatography A 1112: 224–231. II Petsalo A, Turpeinen M & Tolonen A (2007) Identification of bupropion urinary metabolites by liquid chromatography-mass spectrometry. Rapid Communications in Mass Spectrometry 21: 2547–2554. III Tolonen A, Petsalo A, Turpeinen M, Uusitalo J & Pelkonen O (2007) In vitro interaction cocktail assay for nine major cytochrome P450 enzymes with thirteen probe reactions and a single LC/MS/MS run; analytical validation and further testing with monoclonal P450 antibodies. Journal of Mass Spectrometry 42: 960–966. IV Petsalo A, Turpeinen M, Pelkonen O & Tolonen A (2008) Analysis of nine drugs and their cytochrome P450 specific probe metabolites from urine by liquid chromatography-tandem mass spectrometry utilizing sub 2 µm particle size column. Journal of Chromatography A 1215: 107–115. V Tertti K, Petsalo A, Niemi M, Ekblad U, Tolonen A, Rönnemaa T, Turpeinen M, Heikkinen T & Laine K (2011) Transfer of repaglinide in the dually perfused human plancenta and the role of organic anion transporting polypeptides (OATPs). Manuscript.

Some unpublished results are also included.

13 14 Contents

Abstract Tiivistelmä Acknowledgements 9 Abbreviations 11 List of original articles 13 Contents 15 1 Introduction 17 2 Literature review 19 2.1 Plant flavonoid secondary metabolism ...... 19 2.2 Mass spectrometric methods in flavonoid studies ...... 21 2.3 Xenobiotic (drug) metabolism...... 23 2.4 Cytochrome P450 enzymes ...... 25 2.5 LC/MS methods in metabolite profiling and quantitation ...... 27 2.6 Drug interaction studies with N-in-one cocktails ...... 29 3 Aims of the study 33 4 Experimental procedures 35 4.1 Sample preparation ...... 35 4.2 Instrumentation ...... 36 4.3 Matrix effect tests ...... 36 5 Results and Discussion 39 5.1 Structure analysis of flavonoids and drug metabolites [I–II] ...... 39 5.1.1 Flavonoids ...... 39 5.1.2 Bupropion metabolites ...... 44 5.1.3 TOF vs. QqQ ...... 48 5.2 Quantitative analysis of CYP specific probe substrates and their metabolites (in vitro and in vivo) [III - V]...... 50 5.2.1 Method development and validation ...... 50 5.2.2 Comparison of N-in-one assay and single analyte analysis ...... 63 6 Conclusions 65 References 67 Original articles 77

15 16 1 Introduction

Liquid chromatography (LC) combined with mass spectrometry (MS) is a powerful tool in, inter alia, pharmaceutical and plant metabolism analytics, and the use of this hyphenated technique is now commonplace in bioanalytical laboratories [1, 2]. This is a result of recent developments in separation sciences and instrumentations in general, and leads to the use of modern ultra-performance liquid chromatography (UPLC) together with the high selectivity and performance of mass spectrometry. The enhanced separation of analytes in faster analysis time, yielding better throughput with superior results, is leading bioanalytical laboratories to shift from traditional high-performance liquid chromatography (HPLC) to UPLC [3, 4]. Knowledge of the metabolic characteristics of new drug candidates plays a very important role in the selection of suitable compounds for the later stages of drug development projects, and helps reduce the amount of drugs withdrawn from clinical trials and market. The most common reasons for the disqualification of drug candidates may include the formation of reactive metabolites as well as interaction problems directed at the drug metabolizing cytochrome P450 (CYP) enzymes [5, 6]. Careful and comprehensive drug screening and identification of the metabolites thus forms an integral part of the ADMET (absorption, distribution, metabolism, excretion and toxicology) data in the development of drugs. Liver CYP enzyme interaction studies have traditionally been performed separately using specific in vitro assays for each enzyme, utilizing HPLC or LC/MS/MS methods for analysis [7]. However, in recent years, cocktail / N-in-one assays have also been developed for simultaneous screening of several enzyme interactions [8–10] and similar assays have also been developed for in vivo drug-drug interaction studies and CYP activity and phenotyping purposes [11–13]. This thesis describes the utilization of traditional HPLC and improved performance UPLC techniques combined with time-of-flight (TOF) and triple quadrupole (QqQ) mass spectrometry for studies of plant and drug metabolism. Using the developed qualitative and quantitative LC/MS methods, compounds were identified and analyzed from plant extracts and human liver microsome homogenates, urine and placental perfusates. The great variety of sample matrices in the study emphasizes sample preparation with respect to accuracy of the results and thus various methods, such as protein precipitation (PP)

17 and solid phase extraction (SPE), were used to process the biomatrices prior to LC/MS analysis.

18 2 Literature review

2.1 Plant ﬂavonoid secondary metabolism

Flavonoids are polyphenolic compounds with a wide variety of biological activities. They are plant pigments that color most flowers and also protect plants against UV- light and phytopathogens [14, 15]. Their beneficial effects on human health are also widely-known. Flavonoids reduce the risk of cardiovascular diseases [16–18] and cancer [19, 20]. Their ability to act as antioxidants [21–23] and their anti-inflammatory and antimicrobial effects are favorable in the treatment of illnesses such as diabetes and rheumatic diseases [23–25].

The chemical structure of flavonoids is based on the C6 −C3 −C6 skeleton, which is formed by a series of biosynthesis reactions (Figure 1). The main dietary flavonoid subclasses are flavanones, flavones, isoflavones, flavonols, anthocyanidins and flavan-3- ols [26]. The initial material, 4-coumaroyl-CoA, in the flavonoid pathway is derived from phenylalanine via cinnamic acid and p-coumaric acid through the shikimate pathway [27, 28]. The flavonoid biosynthetic pathway starts with the chalcone synthase (CHS) catalyzed reaction of one molecule of 4-coumaroyl-CoA and three molecules of malonyl-CoA. This condensation reaction yields naringenin-chalcone. Chalcone isomerase (CHI) catalyzes the isomerisation reaction from chalcone to flavanone, naringenin. From flavanone the flavonoid pathway diverges into side branches resulting in flavone, isoflavone and dihydroflavonol products. These reactions are catalyzed by flavone synthase (FNS), isoflavone synthase (IFS) and flavanone 3-hydroxylase (F3H) enzymes, respectively. The F3H catalyzed reaction is a stereospecific hydroxylation reaction from flavone to dihydroflavonol. The first step in converting the dihydroflavonol (dihydroquercetin) into anthocyanidin is the reduction of the C-ring of dihydroquercetin with the catalyzing enzyme dihydroflavonol 4-reductase (DFR) [29]. Anthocyanidin synthase (ANS) converts the intermediate leucocyanidin into cyanidin, which in turn is converted to (-)-epicatechin (flavan-3-ol) by anthocyanidin reductase (ANR). Flavonoids are mostly present in plants as glycosides, O-glycosides and C-glycosides. O-glycosides have a sugar moiety bound to the hydroxyl group of the aglycone, and in the case of C-glycosides the sugars are linked to the aglycone with a carbon-carbon bond. The most typical monosaccharides are glucose and rhamnose, but in rare cases

19 Fig 1. Part of the flavonoid pathway and key enzymes involved in the biosynthesis of the flavonoid subclasses. Enzyme abbreviations: CHS, chalcone synthase; CHI, chalcone isomerase; FNS, flavone synthase; IFS; isoflavone synthase; F3H, flavanone 3-hydroxylase; FLS, flavonol synthase; F3’H; flavonol 3’-hydroxylase; DFR, dihydroflavonol 4-reductase; ANS, anthocyanidin synthase; ANR, anthocyanidin reductase [26, 28].

20 arabinose and xylose also appear [30]. Flavonols (kaempferol in the Figure 1) are most widespread of the ﬂavonoids and are found most frequently as O-glycosides, where one or more hydroxyl groups are replaced by a glycosyl group with a C–O glycosidic bond. The glycosylation reaction is catalyzed by glycosyltransferase. The most common hydroxyl groups for glycosylationin ﬂavonols are at the 3- and 7-positions [1].

2.2 Mass spectrometric methods in ﬂavonoid studies

A structure analysis of flavonoids and other plant secondary metabolites is essential when searching for new biologically active compounds. Structural determination is a complicated task, especially when analyzing flavonoids from complex plant matrix extracts. Nuclear magnetic resonance (NMR) and X-ray crystallography are advanced analytical techniques for structure elucidation, but these techniques require a large amount of purified sample isolated from plant extracts. Hyphenated techniques such as liquid chromatography-mass spectrometry (LC/MS) is one useful option for overcoming the problem of a limited amount of sample. A large number of papers concerning the methods for identifying flavonoid structure by tandem mass spectrometric (MS/MS) methods have been published [1, 31–39]. Mass spectrometric methods provide information about flavonoid aglycones and their glycosidic moieties. The nomenclature system for MS fragment ions for flavonoid aglycones and glycosides [40, 41] is widely accepted and used. The nomenclature system for flavonoids is presented in Figure 2. Glycosides can be differentiated by their positive and negative ionization spectra. Positive ionization is more commonly used in the structure analysis of flavonoids [1] although, negative ionization is more sensitive to flavonoids [32, 42, 43]. Fragments i, j containing the aglycone part of the glycoside are labeled Xk,Yk and Zk, the subscript indicating the number of the interglycosidic bond broken (counted from the aglycone), whereas the supercripts of the ion X indicate the cleavages within the sugar rings (Figure 2). Application of low or medium fragmentation energy in collision induced dissociation (CID) for O-glycosides results in the heterolytic cleavage of their glycosidic bonds, + which yields characteristic Yk fragments containing the aglycone part of the molecule [34, 44–46]. Low fragmentation energy does not provide adequate fragmentation for C-glycosides. Medium fragmentation energy, however, causes intraglycosidic cleavages for C-glycosides resulting in i, jX+ fragments that may cause interpretation difficulties

21 Fig 2. Fragment nomenclature used for flavonol glycosides [40, 41]. since the higher fragmentation energy also yields Yk fragments for C-glycosides. Both i, j + + fragments, X and Yk , were detected for C,O-glycosides [47]. The sugar type of the glycan part of the flavonols can easily be determined by the fragments appearing at different m/z values for hexoses (glucose), deoxyhexoses (rhamnose) and pentoses (arabinose/xylose). The characteristic mass losses for the glycan parts of O-glycosides are -162 Da, -146 Da and -132 Da, respectively. Sugar units can be attached directly to the hydroxyl groups in the aglycone or to another sugar moiety. Di-O-glycosides (two sugars at different positions forming monosaccharides) or two sugars at same position (O-diglycosides, forming disaccharide) can be differentiated by their Y1,Y0 and Z1 ions in the product ion spectra. In the case of O-diglycosides, fragmentation of the inner glucose residue is also possible using low-energy CID in the positive ionization mode. The irregular Y∗ ion formed indicates the internal residue loss and may have a very high relative abundance and thus yields false information about the glycan sequence [34]. Loss of inner sugar from the two ring-sequence of O-diglycosides cannot occur in the negative ion mode [1] and thus detection of two independent losses of one sugar moiety from the [M – H] ion may be interpreted as fragmentations from a di-O-glycoside. Radical fragment ion formation has been used for the elucidation of the glycosylation position in the flavonol aglycone [38, 48–52]. The glycan part fragments by homolytic cleavage and forms a radical ion. This homolytic cleavage occurs more easily from the 3-position than the 7-position [48]. The attachment site of the glycan part can be differentiated using the relative abundances of the radical fragment ions and ions from the heterolytic cleavages obtained from the product ion spectra [38, 49]. However,

22 when using homolytic cleavage as a tool for structure elucidation of flavonoids, one should be aware that the instrumentation used [38, 51] as well as the length of the saccharide chain at the 3-position of the flavonol glycosides [52] affect the homo- and heterolytic cleavages. The validation of the fragmentation rules is therefore considered a prerequisite when different types of instruments are used to identify glycosylation sites [51]. The aglycone part of flavonoid O-glycosides can be identified by means of MS3 experiments of the Y0 fragment. Flavonol and flavone aglycones with the same molecular mass can be differentiated from the spectra since the fragmentation pathways are different, corresponding to the substitution pattern of the molecule [40, 43]. The 0,2A+ fragment is charasteristic for flavonols, whereas the abundance of 1,3A+ and 0,2B+ ions is high for both groups of compounds [39, 40].

2.3 Xenobiotic (drug) metabolism

Chemicals foreign to the normal composition of the human body are called xenobiotics. These chemicals include a wide variety of drugs, pollutants, pesticides and food additives. In the human body xenobiotics are transformed into a more hydrophilic form for excretion via kidneys or biliary system. The aim of this metabolism process is to eliminate xenobiotics and to control the levels of the desirable compounds [5]. Biotransformation reactions convert drug molecules to inactive, active and toxic metabolites. In most cases the resulting product of an active substrate is an inactive metabolite. Active metabolites are often formed from an active drug, but in some cases biotransformation reactions convert an inactive substrate to an active metabolite [53]. This phenomenon is utilized, for example, with prodrugs. The metabolic product may also be toxic [54]. An example of a toxic metabolite is N-acetyl-p-benzoquinone imine (Figure 3f), which is a reactive liver toxin resulting from the metabolic modiﬁcation of acetaminophen [55, 56]. Metabolism reactions are usually divided into phase I and phase II reactions [57]. In the phase I a functional group, such as a hydroxyl group, is added to the substance or in some cases the functional group is exposed from the structure by biotransformation reactions [58]. The cytochrome P450 is the most important phase I enzyme system (for more information, see chapter 2.4). The phase II reactions are conjugation reactions, where an endogenic product such as a glucuronic acid, sulfate or glutathione is transferred to the functional group formed in the phase I metabolism.

23 Sometimes the conjugation takes place directly with the unmodiﬁed substrate. The phase II reactions are catalyzed by different conjugative enzymes (Table 1). These metabolites are rarely pharmacologically active.

Table 1. Phase II reactions.

Reaction Enzyme Conjugate Glucuronidation UDP-Glucuronosyltransferase Glucuronide

Sulfation Sulfotransferase -SO3 Glutathione conjugation Glutathione-S-transferase Glu-Cys-Gly

Methylation Methyltransferase -CH2

Acetylation Acetyltransferase -OCCH2 Amino acid conjugation Various Various

Glucuronidation is the most important conjugation reaction in mammals [59]. Glu- curonic acid, originating from UDPGA (uridine-5’-diphosphoglucuronic acid), is conjugated via a glycosidic bond to the substrate. The reaction is catalyzed by UDP- glucuronosyltransferase (UGT). UGTs are a superfamily consisting of a number of isoenzymes [60, 61]. The formed glucuronide is very polar and thus more easily excreted. ROH +UDPGA → RO − GA +UDP (1) Sulfate conjugation is also a signiﬁcant metabolic route in mammals. The sulfate donor in this reaction is PAPS (3’-phosphoadenosine-5’-phosphosulfate) and the reaction is catalyzed by the sulfotransferase enzyme. The sulfate conjugate is anionic and thus much more soluble than the parent compound. In the Equation 2, S=SO3.

ROH + PAPS → RO − S + PAP (2)

Glutathione (GSH) is a tripeptide formed from glutamate, cysteine and glycine. The glutathione conjugation occurs easily with electrophilic groups and is an important metabolic route, especially for reactive metabolites. It is a defence mechanism against carcinogens. Glutathione binds to substrate through the cysteine nucleophilic thiol group. RH + GSH → R − S − G (3) The conjugate formed is modiﬁed further as the glutamyl and glycinyl groups are removed while the remaining cysteinyl group is acetylated. The product of the yield, a

24 mercapturic acid derivative (=N-acetyl-cysteine conjugate), is excreted into the urine. Other phase II reactions are methylation and acetylation. These reactions do not improve solubility but rather reduce pharmacological activity [57]. Methylation is not very common for exogenous drug molecules but it is a conventional metabolism route for endogenic substances. The methyl group from the cofactor S-adenosylmethionine (SAM) is transferred to the substrate by methyltransferases. In acetylation the acetyl group donor is Acetyl-CoA and the reaction is catalyzed by N-acetyltransferase (NAT).

X + AcetylCoA → Acetyl − X +CoA (4)

Amino acid conjugations are also possible phase II metabolism reactions, in particular with the carboxylic acids [5]. The amino acid in the conjugation reactions is usually glutamine or glycine. The formation of hippuric acid through the glycine conjugation of benzoic acid is an example of amino acid conjugations. The metabolism of drugs and other xenobiotics is affected by matters such as genetic, non-genetic and environmental factors [62, 63], which lead to a wide inter- individual variability of drug metabolism. The most important exogenous factors affecting xenobiotic metabolism are substances with inhibitive or inductive effects towards drug metabolizing enzymes. Polycyclic aromatic hydrocarbons (PAH) are one of the best known substances with inductive properties. Hypericum perforatum (St. John’s wort) also has drug interaction tendencies through the induction of CYP3A4 and CYP2C9 enzymes [64, 65]. In addition, many drugs, for example, ketoconazole (CYP3A4), ﬂuvoxamine (CYP1A2) and quinidine (CYP2D6), are known to be inhibitors [6, 66, 67].

2.4 Cytochrome P450 enzymes

The cytochrome P450 (CYP) enzymes are a superfamily of hemoproteins which catalyze a wide range of biotransformation reactions. The CYP enzymes convert foreign substances into a more suitable, i.e. hydrophilic, form for excretion. The cytochrome P450 enzymes are undoubtedly the most important metabolizing enzyme system in humans since the CYP enzymes are responsible for approximately 70-80% of the phase I (oxidative) metabolism of drugs and other xenobiotics [68]. The cytochrome P450 enzymes are distributed in almost every human organ, but the highest concentrations of CYP enzymes are found in the liver. The liver is the center of xenobiotic metabolism and is capable of catalyzing both oxidation and reduction pathways [69]. Cytochrome

25 P450-catalyzed reactions can be broadly classiﬁed into four categories: hydroxylations, epoxidations, reductions and heteroatom oxidations [70]. Examples of some CYP catalyzed phase I reactions are shown in Figure 3.

Fig 3. Some CYP catalyzed phase I reactions [57, 62].

The CYP metabolic system consist of the CYP enzyme, molecular oxygen and NADPH- cytochrome reductase, which serve to transfer electrons. The generic oxidative reaction for a given substrate R may be presented as follows:

+ + RH + O2 + NADPH + H → ROH + NADP + H2O (5)

The cytochrome P450 superfamily is divided into a number of families and subfamilies according to the amino acid composition of the enzymes. Genes in the same subfamily have a similarity greater than 55% in their structures [5]. The CYP1-, CYP2- and CYP3 families are mainly responsible for xenobiotic metabolism; the most important enzymes are CYP1A1/2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1 and CYP3A4 [71]. In the human liver the most important drug metabolizing enzyme is CYP3A4, whereas the members of other CYP families than mentioned here are largerly

26 responsible for the biotransformation of endogenic components such as steroids and fatty acids [72].

2.5 LC/MS methods in metabolite proﬁling and quantitation

Nowadays LC/MS techniques have clearly become the most widely-used tool in analyzing drugs and their metabolites [2, 73]. Typical steps in these analyses consist of sample preparation, chromatographic separation, mass spectrometric detection and data processing. The sample preparation is an important step in all analyses, and especially when analyzing biological fluids [3, 74]. The analyte must be extracted from the sample matrix components and other impurities that may reduce the specificity of detection or interfere with the mass spectrometric ionization process by suppression or enhancement. A commonly used sample preparation method in LC/MS analysis is still traditional liquid-liquid extraction (LLE), although the method is time-consuming and produces great amounts of solvent waste [3]. Another commonly used method, solid phase extraction (SPE), uses a sorbent material containing a solid phase and a flowing liquid phase to isolate a compound or certain type of compounds from a liquid sample. The sample is loaded into cartridge, undesired compounds are washed away, and the analytes are eluted with a different solvent [75]. Numerous SPE cartridges containing different packing materials are available and can be used for extracting acidic, basic and neutral compounds from different kind of matrices and also use the ion exchange principle [76]. It is worth noting that when using extraction methods the discarding of metabolites with different recoveries can occur and thus modify the metabolic profile of the study compound. Therefore, in the case of metabolite profiling, it is recommended that unspecific sample preparation methods such as protein precipitation with acetonitrile or methanol rather than SPE or LLE be used. Chromatographic separations are most commonly obtained using high-performance liquid chromatography (HPLC) with reversed-phase columns. In the early phase of metabolite profiling rapid generic chromatographic methods are often utilized, applying 5-80% acetonitrile or methanol and aqueous phase in a gradient run of some minutes. The optimization of gradient strength and aqueous phase pH using parent compound, however, also improve peak shapes and thus lead to better chromatographic resolution

27 and detection sensitivity [73]. As column properties strongly affect separation efficiency, it is most convenient to use short columns with a narrow inner diameter (2 mm x 50 mm) packed with small particles (≤ 3 µm). Techniques using columns with a particle size of sub-2 µm and an instrumentation capable of working with elevated back pressures are usually called ultra-performance liquid chromatography (UPLC) [77–80]. The smaller particle size enables a shorter analysis time and better resolution. For metabolic studies the most common ionization methods in LC/MS applications are electrospray ionization (ESI) and atmospheric pressure chemical ionization (APCI). The ion sources are referred to as API sources, as the ionization does not take place in a vacuum but in atmospheric pressure. ESI is suitable for almost all drug-like molecules with at least one easily ionizable functional group [81], whereas for steroids and other less polar compounds, APCI is often utilized [82]. Of these two methods, APCI is generally less susceptible to matrix effects [83, 84]. The negative ionization polarity is considered more specific, whereas the positive ionization mode is more vulnerable to ion suppression [84, 85]. In ESI, the analyte is ionized within the solution and thus the chemistry of the liquid phase has a great impact to ionization efficiency. The solution containing the analyte is directed through the high voltage capillary (2 - 5 kV), where the capillary voltage gathers the oppositely charged ions to the sides of the capillary and thus the droplets spraying from the capillary are excessively charged with the same polarity as that in the capillary. The liquid flow and gas flow parallel to the capillary aid the formation of droplet mist. The droplets diminish by means of vaporization (aided by drying gas flow) until the repulsion forces of the similarly charged ions are larger than the cohesive strengths of the surface tension, when the individual ions vaporize to gas phase (ion evaporation method, IEM). In the charge residue model (CRM) the vaporization of the solvent continues until only single ion droplets remain. The IEM is believed to be dominant with the small molecules, whereas in the case of large molecules, such as proteins, CRM occurs [86]. The ion suppression / enhancement by the co-eluting component may occur due its effect to droplet formation or the availability of charge (proton), or co-precipitation of the analyte from the spray as a neutral molecule [87–89]. From the MS analyzer point of view, quadrupole instruments are most used in quantitative analysis due to their high sensitivity, good specificity and high linear range, whereas TOF instruments are used in screening type analyses, such as profiling of unknown metabolites, because they offer high sensitivity wide mass range scan with high mass resolution (>40 000 FWHM) and high mass accuracy. However, modern

28 time-of-flight mass spectrometers could be used for quantitative work as their linear range has significantly improved recently due to the introduction of dynamic range enhancement systems [90]. TOF instruments are capable to very high data acquisition rates and thus ideal for UPLC. Ion trap mass spectrometers are most commonly used for screening type analysis because of their wide mass range and relatively high sensitivity [91] and capability to MSn in structure characterization, but their use with UPLC is limited by their relatively slow data acquisition rate. MS data processing is nowadays highly automated and the development of post- acquisition data mining techniques are strongly enabled by high mass resolution data. Metabolite screening is assisted by software that compares the acquired data between the sample and negative control and suggests biotransformations and calculated elemental compositions for the found metabolites using accurate mass measurements and information about parent compound. Some software includes features like the dealkylation tool [92] and are capable of fragment ion data interpretation [93, 94] and mass defect filtering (MDF) [95, 96]. Fragment ion data is used to identify biotransformation sites in the parent molecule by the use of computer algorithms that compare mass shifts of the diagnostic fragment ions to the parent molecule and detected metabolites. A dealkylation tool is used to identify bonds having the potential to cleave by biotransformation reactions, whereas MDF removes biological background using the observation that biotransformations usually change relatively little the decimal component of the m/z observed [92]. In addition, using current in silico methods together with accurate mass measurements for predicting metabolism and mass fragmentation is a solution to differentiate parent molecules and metabolites with identical molecular formulae when there is an absence of reference standards or spectra [97].

2.6 Drug interaction studies with N-in-one cocktails

Drug-induced inhibition of CYP enzymes is known to be the most common reason for drug-drug interactions and therefore studying the inhibitory effect of new chemical entities is extremely important when selecting lead compounds and for minimizing the number of drugs being withdrawn from clinical studies [6, 98, 99]. N-in-one assays can be used to study a number of different CYP enzymes at the same time, increasing the amount of data obtained from a single experiment, and thus increasing throughput and saving time. The three to six most usually included CYP isoenzymes in the N-in-one assays are CYP1A2, CYP2C9, CYP2C19, CYP2D6, CYP2E1 and

29 CYP3A4. Generally, the cocktails lack the probe metabolites for isoenzymes CYP2A6, CYP2B6 and CYP2C8. A number of papers describing analytical methods for monitoring several CYP- specific probe metabolites from human urine or plasma after a single or N-in-one administration ("cocktail/cassette dosing") have been published for in vivo interaction studies or cytochrome P450 activity and phenotyping purposes [11, 12, 100–112]. The cytochrome P450 interaction studies have been traditionally conducted with separate specific assays for each single CYP isoform, which offers the possibility to optimize the analysis for only one compound at a time. This provides the best possible performance for the analysis. For example, the effect of HPLC eluent pH can be optimized for the best chromatographic peak shape and for optimum detection response, i.e. for mass spectrometric ionization. The requirements for the analytical method with N-in-one assay are clearly of a higher level, as it is often the case that many substances with chemically very diverse properties (acids or bases with high or low hydrophilicity) should be analyzed as fast as possible, and preferably from a single analysis from a single sample. Therefore the set-up of a functional analytical method is most often the search for the best compromise in the analytical performance between the completely different analytical conditions, which are optimum for each compound. Thus far, almost all published in vitro N-in-one assays for CYP activity screening have relied on reversed-phase liquid chromatography (RP-HPLC) as an analytical method, and most commonly with mass spectrometric detection [8, 9, 13, 113–118]. Only few in vitro and in vivo N-in-one assays have utilized ultra performance liquid chromatography [10, 111, 119]. The need for throughput is high in the drug industry due to the large amount of lead compounds produced by combinatorial chemistry and the samples obtained from metabolism studies. Analysis time is thus an important factor in CYP interaction screening. In the published assays analysis time (gradient elution and column equilibration) varies from less than 1 min to 15 min cycle time. The time used for each individual sample is not only affected by sample preparation and chromatographic methods, but also by the mass analyzer used. LC/MS detection with triple quadrupole mass spectrometers is usually preferred over other types of instruments not only because of their best linear range of detection, but also due to the highly specific multiple reaction monitoring (MRM, sometimes also called selective reaction monitoring, SRM). In MRM, two mass filters are used for ion selection; first the precursor (parent) ion is selected (Q1) to collision cell (q2) and, after the collision with gas atoms (CID, collision

30 induced dissociation), the speciﬁc fragment ions of the parent are selected with the second mass ﬁlter (Q3). The use of MRM enables fast chromatrographic runs due to a reduced need for complete chromatographic separation for most of the analytes.

31 32 3 Aims of the study

The aim of the study was to develop faster and more sensitive LC/MS techniques for the identiﬁcation and quantitation of plant and drug metabolism products. A number of cocktail-type N-in-one-assays have been developed for screening CYP inhibition or induction for different CYP enzymes. These assays have, unfortunately, not been extensive enough or the analysis methods used have been laborious to perform. Therefore, the main focus was in developing LC/MS analysis methods for CYP-speciﬁc drug interaction studies, i.e. for N-in-one in vitro and in vivo cocktail assays. The detailed aims of the study were

– to identify flavonoids of Rhodiola rosea [I] by liquid chromatography-mass spectrometry and to test the suitability of collision-induced radical cleavage of flavonoid glycosides as a tool for glycosylation site identification – to identify LC/MS urinary metabolites of bupropion [II] for the development of an in vivo cocktail assay – to develop and validate LC/MS/MS analysis methods for an in vitro interaction cocktail assay [III] and for nine drugs and their cytochrome P450-specific probe metabolites from human urine [IV] – to develop a specific and sensitive analysis method for the quantitation of repaglinide from human placental perfusates [V]

33 34 4 Experimental procedures

This section brieﬂy presents the sample preparation methods, instrumentation and matrix effect tests used in this study. More detailed descriptions of the sample preparation methods, experimental parameters, columns and additional instruments can be found in original publications I–V.

4.1 Sample preparation

The sample matrices used in this study were dried plant material, urine and samples from liver microsomal incubations and placenta perfusions. The plant material used for Rhodiola rosea flavonoid analysis in study I was material grown at the Department of Biology and Botanical Gardens at the University of Oulu. The sample preparation methods used for the plant material were supercritical fluid extraction (SFE) and liquid extraction with aqueous methanol in an ultrasonic bath. Human urine samples in studies II and IV were incubated in the presence of β- glucuronidase for the hydrolysis of glucuronide conjugates of the phase I metabolites. In addition, solid phase extraction with different sorbents was tested on the urine samples (unpublished data); the sorbents were Oasis HLB (macroporous copolymer, 10mg), MAX (strong anion exchange, 30 mg), MCX (strong cation exchange, 30 mg), WAX (weak anion exchange, 30 mg) and WCX (weak cation exchange, 30 mg) from Waters (Waters Corp., Milford, MA, USA). The urine sample was loaded into the cartridge without any pH modifications after glucuronide hydrolysis. The wash and elution steps were performed according to the manufacturer’s standard instruction protocol as follows: MCX/WAX and WCX/MAX cartridges were washed with 2% formic acid in water and 5% ammonium hydroxide in water, respectively, while the HLB cartridge was washed with 5% aqueous methanol. Acetonitrile:methanol (1:1) mixture was used for the elution of analytes from HLB cartridge, and methanol was used for the first elution (wash step 2) with the MCX/WAX and WCX/MAX cartridges. For the final elution of the analytes from MCX/WAX and WCX/MAX cartridges, 5% ammonium hydroxide in methanol and 2% formic acid in methanol were used, respectively. After SPE, samples were evaporated to dryness and reconstituted to 300 µl 1:5 mixture of acetonitrile and phosphate buffered saline (PBS), which was used for retaining good peak shape. The

35 trial for dosing and collecting human urine samples was designed and monitored in accordance with Good Clinical Practice and the Declaration of Helsinki. The human in vitro liver microsomal incubations in study III were performed as described by Turpeinen et al. [114] using 0.5 mg/ml microsomal protein content and NADPH as a cofactor. In the repaglinide study V, solid phase extraction using Oasis MAX anion-exhange 96-well plates was utilized for the placenta perfusion samples, and for antipyrine analysis, the samples were used as such, without any pre-treatment.

4.2 Instrumentation

The LC/MS instrumentation used in this study is presented in Table 2. The liquid chromatographic separations were performed using the HPLC and UPLC systems. The initial screening and accurate mass measurements of the compounds and metabolites present in the sample matrices were carried out using an LCT TOF mass spectrometer equipped with a LockSpray ion source. The more detailed structure determinations and quantitative analysis were performed using triple quadrupole mass spectrometers. All three Waters Quattro mass spectrometers were equipped with Z-spray ion sources. Electrospray ionization with positive and negative polarities was used in all studies. Several chromatographic columns and mobile phase eluents were tested when developing the LC/MS methods used in the studies. The LC columns and eluents chosen for the studies are presented in Table 3.

4.3 Matrix effect tests

The effects of matrix components on the ionization response of analytes were tested by two different test approaches [89, 120–122]: a) comparing the LC/MS/MS peak intensity obtained from samples spiked to low analyte concentration (ng/ml) to blank sample matrix and blank solvent (sample matrices processed similar to the real samples) and b) injecting a blank matrix sample (processed similar to the real samples) and monitoring the suppression or enhancement of the baseline for the analyte compounds achieved by post-column infusion of the analyte compound solution into the ﬂow directed to the ion source.

36 Table 2. LC/MS instruments used in studies I–V.

Liquid chromatographs Mass spectrometers Waters 2690 Alliance HPLC system (Waters Cor- Micromass Quattro II triple quadrupole mass spec- poration, Milford, USA) [I] trometer (Micromass, Altrincham, UK) [I] LCT TOF time-of-ﬂight mass spectrometer with LockSpray ion source (Micromass, Altrincham, UK) [I, II, IV] Waters 2695 Alliance HPLC system (Waters Cor- Micromass Quattro Micro triple quadrupole mass poration, Milford, USA) [II-IV] spectrometer (Micromass, Altricham, UK) [II-IV] Finnigan Surveyor HPLC (Thermo Finnigan, San TSQ Quantum Discovery MAX triple quadrupole Jose, USA) [III] mass spectrometer (Thermo Finnigan, San Jose, USA) [III] Waters Acquity UPLC system (Waters Corpora- Micromass Quattro Premier triple quadrupole tion, Milford, USA) [IV, V] mass spectrometer (Micromass, Altrincham, UK) [IV, V]

Table 3. LC columns and mobile phase eluents used in the studies I–V.

Study Columns I Waters Symmetry C18 (2.1 mm × 100 mm, 3.5 µm) II Phenomenex Gemini C18 (2.0 mm × 150 mm, 5.0 µm) III Waters Sunﬁre RP18 (2.1 mm × 100 mm, 5.0 µm) IV Waters BEH Shield RP18 (2.1 mm × 50 mm, 1.7 µm) (UPLC) Waters XBridge Shield RP18 (2.1 mm × 50 mm, 3.5 µm) V Waters BEH Shield RP18 (2.1 mm × 50 mm, 1.7 µm) (UPLC) Study Mobile phase eluents I A) 0.1 % formic acid B) acetonitrile II A) 0.1 % acetic acid (ESI+), 2 mM ammonium acetate (ESI−) B) methanol III A) 1% formic acid + 10 mM ammonium acetate (pH 2.4) B) methanol IV A) 0.1 % acetic acid B) acetonitrile V A) 0.1 % acetic acid B) acetonitrile

37 38 5 Results and Discussion

The main results of the study are described and discussed in this section. Some unpublished data is also presented in the method development section (5.2.1). More detailed information can be found in the original publications I-V.

5.1 Structure analysis of ﬂavonoids and drug metabolites [I–II]

5.1.1 Flavonoids

LC/TOF-MS was used for screening flavonoids from the Rhodiola rosea extracts. A total of 15 flavonol glycosides were detected by LC/MS from the ethanol extracts of the aerial parts of the R. rosea. The LC/TOF-MS chromatograms of the detected flavonoids are presented in Figure 4 and the molecular weights and retention times collected from the analysis together with the identifications of the detected flavonoids are shown in Table 4; the structures of all the detected flavonoids are presented in Figure 6. The initial identifications of the flavonoids were obtained using accurate mass measurements and detailed identifications were carried out using tandem mass spectrometric methods. Five of the detected flavonoids (7, 9, 11, 13 and 15) were known from the literature [123–126] and the remainder (1-6, 8, 10, 12 and 14) were previously unreported from Rhodiola rosea. Most of the flavonols were detected with both positive and negative electrospray ionization (Figure 4); flavonols 9 and 10 were detected using negative ionization mode since negative electrospray ionization is more sensitive to flavonoids [32, 42]. The in-source generated aglycone [M + H–glycose]+ fragment-ions were chosen for further MS/MS experiments ( = "pseudo MS3") to identify the aglycone parts of the detected flavonoids. All aglycones were identified by comparing their spectra with those of standard compounds and the literature [36, 127]. The data from the "pseudo MS3" experiments is presented in Table 3 in original publication I. The aglycones were identified as kaempferol, herbacetin, gossypetin, quercetin and hydroxyl-gossypetin, in which the hydroxylation sites in the B-ring were identified as propably being 3’-, 4’- and

39 5’-hydroxylated, as in ﬂavonoid hibiscetin.

Table 4. Detected ﬂavonoids in R. rosea extracts.

Flavonoid RT (min) MW Identiﬁcation 1 11.8 642 Gossypetin-di-O-glucoside 2 13.4 642 OH-gossypetin-7-O-rha-8-O-glu 3 14.3 626 Herbacetin-di-O-glucoside 4 15.1 610 Kaempferol-3-O-glu-7-O-glu 5 15.1 610 Quercetin-3-O-rha-7-O-glu 6 15.5 642 Gossypetin-di-O-glucoside or O-di-glucoside 7 16.3 626 Gossypetin-7-O-rha-8-O-glu = rhodiolgidin 8 16.5 612 Gossypetin-3-O-glu-7-O-xylo/ara 9 17.6 596 Herbacetin-8-O-xylo-3-O-glu = rhodalidin 10 17.6 594 Kaempferol-3O-rha-7-O-glu 11 18.7 610 Herbacetin-7-O-rha-8-O-glu = rhodionidin 12 18.8 596 Herbacetin-3-O-glu-7-O-xylo/ara 13 27.8 464 Gossypetin-7-O-rha = rhodiolgin 14 31.0 448 Quercetin-3’/4’-rha 15 32.4 448 Herbacetin-7-O-rha = rhodionin ara: arabinose; glu: glucose; rha: rhamnose; xylo: xylose

The glycosylation sites of the detected flavonols were tentatively identified by comparing the intensities of the CID fragment ions from [M + H]+ and according to the radical fragment ion intensities produced in negative electrospray ionization. The glycoses from different sites of the protonated flavonol di-O-glycosides cleave more easily than others, producing a more intense fragment peak, the cleaving order being: 5 > 3 > •− 3’ ≥ 5’ > 4’ > 7 [1]. Yet, the intensity of radical aglycone [Y0 – H] ion, formed by homolytic cleavage of glycose from the aglycone in the MS/MS of [M – H]− ions is dependent on the position in which the cleaving glycoside is attached to the aglycone; the phenomenon was also tentatively discussed by Hvattum & Ekeberg [48] and Cuyckens & Claeys [38] at the very time we were studying the issue. However, the interpretation of the data should be carried out very carefully, and only after comparing the data with fragmentations of known standard flavonoids obtained with the very same instrument and the same parameters. Therefore, we used six different flavonol

40 Fig 4. LC/MS sum ion chromatograms of detected flavonoids acquired at ESI+ (a) and ESI (b) from the ethanol extract of R. rosea aerial parts. The chromatograms are amplified 24× 1) 11.8 min Gossypetin-di-O-glucoside, 2) 13.4 min OH-gossypetin-7-O-rha-8-O-glu, 3) 14.3 min Herbacetin-di-O-glucoside, 4) 15.1min Kaempferol-3-O-glu-7-O-glu, 5) 15.1 min Quercetin-3-O-rha-7-O-glu, 6) 15.5 min Gossypetin-di-O-glucoside or O-di-glucoside, 7) 16.3 min Gossypetin- 7-O-rha-8-O-glu = rhodiolgidin, 8) 16.5 min Gossypetin-3-O-glu-7-O-xylo/ara, 9) 17.6 min Herbacetin-8-O-xylo-3-O-glu = rhodalidin, 10) 17.6 min Kaempferol-3O- rha-7-O-glu, 11) 18.7 min Herbacetin-7-O-rha-8-O-glu = rhodionidin, 12) 18.8 min Herbacetin-3-O-glu-7-O-xylo/ara, 13) 27.8 min Gossypetin-7-O-rha = rhodiolgin, 14) 31.0 min Quercetin-3’/4’-rha, 15) 32.4 min Herbacetin-7-O-rha = rhodionin [I]. glycosides with glycosylation sites in the 3-, 7-, 8- and 4’-positions for the comparison experiments (Table 1 in original publication I). Three of the flavonol standards used were mono-O-glycosides, two were O-diglycosides (the saccharides were rutinose and neohesperidose) and one was di-O-glycoside, with robinin and rhamnose as the sugar units. The aglycone fragment ions obtained from the standard compounds, as well as the fragment ions due to losses of each glycose moiety from robinin, are presented in Figure 5. When comparing the losses of glycosides from the 3-position (Figure 5a and 5b) and the loss of glucose in the 4’-position (Figure 5c), the cleavages of the glycosides from • the 3-position produced very intense [Y0 – H] ions, as was reported in earlier studies [38, 48], and thus were distinguishable from the cleavage from the 4’-position. The

41 cleavages from the 7- and 8-positions (Figure 5d and 5e) were also clearly distinguishable •− from 3- and 4’-positions, and the relative intensities between the produced [Y0 – H] − ions compared to the [Y0 ] ions were much lower than the intensities produced by the cleavages from the 3- and 4’-positions. However, it was not possible to distinguish between the fragmentation of ﬂavonols glycosylated in 7- and in 8-positions. The cleavage of rhamnose from the 7-position of robinin did not produce a radical ion, but only a [M–H–146]− ion at m/z 593, whereas the loss of robinose (6-rhamnosyl-galactose) from the 3-position produced a very intense radical fragment ion at m/z 430 (Figure 5f). This known phenomenon of an intense radical ion produced from the 3-position of ﬂavonol glycosides has also been corroborated in more recent studies [50, 52].

•− − − Fig 5. [Y0 – H] and [Y0 ] ions from MS/MS spectra of [M – H] acquired from a) quercitrin (rhamnoside in 3-position); b) rutin (rutinoside in 3-position); c) spi- raeoside (glucose in 4’ -position); d) kaempferol-7-neohesperidose; e) gossypin (glucose in 8-position); and f) [M – H – rha]− and [M – H – rob]•− ions acquired from robinin (rhamnose in 7-position and robinose in 3-position). Radical fragment ions marked with circles [I].

42 Fig 6. Flavonoids identiﬁed from the R. rosea extract [I].

43 The experiments for the two known 7-O-rhamnopyranose-8-O-glucopyranosides from the R. rosea, rhodiolgidin (7) and rhodionidin (11), showed that the presence of another glycan moiety in the 7-position (here, rhamnose) strongly increases radical fragment ion formation in the cleavage from the 8-position (Figure 3 and Table 2 in I). Thus, the differentiation of 7-,8-substituted and 3-,7-substituted flavonoid di-O- glycosides is very difficult, the radical ion formation due to cleavage from the 8-position being very similar to that from the 3-position. In addition, the fragmentation behaviour of glycans from the 3- and 8-position in these kinds of di-O-glycosides is also very much alike with positive ionization (Table 2 in I), making the identification of the 3- and 8-positions even more difficult. However, if the aglycone of an unknown flavonoid glycoside does not have hydroxyl groups in both the 7- and 8-positions, the suggestive identification of glycosylation sites may be obtained with correct instrument adjustments and comparison with known standards. After identifying the correct aglycones, compounds 9 (rhodalidin), 13 (rhodiolgin) and 15 (rhodionin), in addition to rhodiolgidin (7) and rhodionidin (11), were identified as the known flavonoids of the plant. The other detected flavonoids were identified using the methods described above (data presented in Tables 2 and 3 in I). The results of this study suggest that radical ion fragmentation could be used as a structure elucidation tool for flavonol glycosides, although, the obtained data should be analyzed very carefully and always in comparison with known standards. This study is, to our knowledge, the only one containing the homolytic fragmentation behaviour of flavonol 8-O-glycosides. It was also the first to report the identification of compounds 1-6, 8, 10, 12, and 14 for Rhodiola rosea.

5.1.2 Bupropion metabolites

Drug metabolites may have some structural features which may enable the formation of reactive species in the human body; thus the structure elucidation of even minor metabolites is important from toxicological point of view. In this study, 20 bupropion metabolites were detected by LC/TOF-MS from human urine samples. Sixteen of the detected metabolites were conjugates, including 12 glucuronides, three sulfates and a glycine conjugate; the other four metabolites were phase I metabolic products. Biotransformations were screened using accurate mass measurements and the detailed identiﬁcation of the detected metabolites was performed using tandem mass spectrometric methods (data shown in Table 1 in II). The LC/TOF-MS chromatograms of the

44 detected metabolites are presented in the Figures 7 and 8. The tandem mass spectrometric experiments confirmed the identifications of hydroxybupropion (M1), threo- and erythrohydrobupropion (M2 and M3) and m-chlorohippuric acid (M5). The M2 was identified as threohydrobupropion as its concentration is known to be much higher than the other detected amino alcohol metabolite, M3 [128, 129]. The M4 was also identified as the previously known metabolite with butyl-group hydroxylation and keto-group reduction to an alcohol [130, 131]. Figure 8 shows the presence of four isobaric glucuronide conjugates for threo- and erythrohydrobupropion (a, M6 - M9), as well as four glucuronidated hydroxymetabolites (b, M10 - M13), three glucuronides for a metabolite with hydroxylation and hydrogenation (c, M14 - M16) and a glucuronide for dihydroxylated bupropion (d, M17), in addition to sulfate conjugates for hydroxylation metabolite (f, M18) and for hydroxylations with hydrogenation (e, M19 and M20). The sulfates were detected only in negative ion mode, whereas all the other metabolites were detected in positive ion mode. The high stability of M6 - M9 to β-glucuronidase activity suggested N-glucuronidation. As the N-glucuronidation may form a new chiral center in the molecule and lead to the formation of a total of four glucuronide-conjugated amino alcohol metabolites, metabolites there were assigned as N-glucuronides of threo- / and erythrohydrobupropion. The configuration of nitrogen can be tetrahedral and the lone nitrogen electron pair can be considered as the fourth group providing chirality, especially if racemization by pyramidal inversion is restricted by glucuronidation. The M12 and M13 (glucuronidations of hydroxybupropion) and M17 (glucuronide of dihydroxylation metabolite) were also resistant to hydrolysis and thus suggest N-glucuronidation. In addition, the formation of two diastereoisomers of the glucuronides M12 and M13 also pointed to N-glucuronidation. In the case of M10, M11 and M14 - M16, the glucuronides were thought to be O-glucuronidations as the hydrolysis rates obtained were similar to the O-glucuronide standards of 6-OH-chlorzoxazone and 1-OH-midazolam. The aglycone parts of the conjugated metabolites were identified by pseudo-MS3 experiments where the in-source generated [M + H –glucuronide]+ and [M – H –sulfate]− fragments were chosen for the collision cell CID. The hydroxylation sites of M10 - M13 were identified by comparing the fragmentation patterns of the aglycone part of the metabolite with those of bupropion and hydroxybupropion (Figure 6 in II). The M10 and M11 were identified as glucuronides of aromatically hydroxylated stereoisomers, whereas the aglycone parts of M12 and M13 were identified as morpholino hydroxy-

45 Fig 7. LC/TOFMS ion chromatograms for bupropion and its phase I metabolites, generated using [M + H]+ ions: a) bupropion; b) threo- / erythrohydrobupropion (M2 and M3); c) hydroxybupropion (M1); d) hydroxylation with hydrogenation (M4); and e) m-chlorohippuric acid (M5) [II].

46 Fig 8. LC/TOF-MS ion chromatograms for conjugated bupropion metabolites, generated using the [M + H]+ and [M – H]− ions: a) [M + H]+ glucuronides of threo- / erythrohydrobupropion (M6 - M9); b) [M + H]+ glucuronides of hydroxylated metabolite (M10 - M13 ); c) [M + H]+ glucuronides of hydroxylation with hydrogenation (M14 - M16). Peaks at retention times of 7.9 min, 9.9 min, 11.2 min and 11.6 min are due to 37Cl isotope of glucuronide conjugates of hydroxymetabolites (m/z 432); d) [M + H]+ glucuronide of dihydroxylation (M17); e) [M – H]− sulfates of hydroxylation with hydrogenation (M19 and M20) and f) [M – H]− sulfate of hydroxylation (M18) [II].

47 bupropion. The cleavage of the butyl-group from the glucuronide conjugate (Figure 9c) at m/z 376 and the fragment at m/z 200 indicate the loss of an intact butyl group from the protonated aglycone part of the molecule. The fragment at m/z 200 is analogous to the ion at m/z 184 rising from the loss of the butyl-group in bupropion and morpholino bupropion (Figure 6a and 6b in II). The fragment at m/z 216 in the spectrum (Figure 9g) of the glucuronide conjugate of the dihydroxylation metabolite (M17) is again analogous to the ion at m/z 184 of the glucuronide conjugate of hydroxybupropion (Figure 9d), indicating that the hydroxylation was not located in the butyl group of the substrate, as was previously reported for the dihydroxymetabolite [132]. The aglycone part of the sulfate conjugate M18 was identified to be the metabolite with aromatic hydroxylation and the aglycone of M19 and M20 was identified as a hydroxylated and hydrogenated metabolite. The structures of the detected bupropion metabolites are presented in Figure 4 in II. In this study, one previously unreported new phase I metabolite, dihydroxylation, was detected. In addition, previously unpublished glucuronide conjugates for dihydroxy- bupropion, hydroxylation with hydrogenation and for two different hydroxy metabolites, together with the sulfate conjugates of aromatic hydroxylation and hydroxylation with hydrogenation, were reported for the first time.

5.1.3 TOF vs. QqQ

When comparing the time-of-flight and triple quadrupole mass spectrometers, the TOF instrument proved to be a powerful tool for screening purposes; the QqQ instrument was used for more detailed elucidation tasks due to a lack of Q-TOF-instrument; TOF may, however, yield very similar results to triple quadrupole MS/MS when using high cone voltages (in-source CID). TOFs offer excellent sensitivity in the high mass range detection mode with very good resolution and mass accuracy. This enables accurate mass measurement and thus identification of the metabolic biotransformation with respect to the parent compounds. Although the triple quadrupole instruments operating with different scan modes, such as neutral loss mode, could be applied to screen flavonol glycosides and conjugation metabolites from plant extracts and in vitro / in vivo samples, respectively, the detection probability of unexpected metabolites is far better with TOF than with triple quadrupole instruments.

48 Fig 9. MS/MS spectra for glucuronide conjugates of a) threohydrobupropion, (M2); b) erythrohydrobupropion, (M3); c) aromatic hydroxylation, (M11); d) hydroxybupropion, (M13); e) hydroxylation with hydrogenation, (M14); f) hydroxylation with hydrogenation, (M16); and g) dihydroxylation, (M17) [II].

49 5.2 Quantitative analysis of CYP speciﬁc probe substrates and their metabolites (in vitro and in vivo) [III - V]

5.2.1 Method development and validation

N-in-one cocktail assay [III, IV]

In recent years, several in vitro and in vivo cocktail / N-in-one assays have been developed for CYP enzyme interaction screening purposes. These assays contain up to seven most important CYP isoforms, but usually exclude the probe metabolites for the isoenzymes CYP2A6, CYP2B6 and CYP2C8, and thus the activity of all drug metabolizing enzymes is not studied [8, 12, 108, 133]. In addition, the analytical method performances in these assays have been somewhat deficient, low sensitivity leading to a need for high parent drug dose levels, which is a drawback when dosing multiple drugs at the same time. In some cases more than one analysis was also required to detect all the metabolites due to the analysis of CYP2E1 probe drug chlorzoxazone [116, 118, 134] or the need to analyze both urine and plasma samples [12, 135]. Here an LC/MS/MS method was developed [III] for our in vitro cocktail assay to extent the linearity and enhance the sensitivity of the earlier LC/TOF-MS method [114]. The in vitro cocktail consisted of ten CYP-selective probe substrates, which lead to the analysis of 13 probe reactions in a single run. In addition, an UPLC/MS/MS analysis method was developed for the in vivo N-in-one cocktail assay [IV] that consisted of nine drugs and their CYP-specific metabolites and utilized a non-invasive sample collection from the patients. The sample matrices used in the N-in-one cocktail assays enabled simple sample preparation steps before LC/MS analysis, and thus the method development in these studies mainly focused on chromatography and mass spectrometry. Only centrifugation was needed for the incubation samples [III] prior to analysis and for the urine samples in the study IV the dilution of the matrix (hydrolysis of the glucuronide conjugates) was considered sufficient to keep the matrix effect at an acceptable level. Solid phase extraction, however, was tested for the N-in-one assay urine samples (subsection below). In both N-in-one cocktail assays, a number of probe metabolites, with a great variation in structures (Figure 10 shows the structures of the in vivo cocktail compounds, IV) and chemical properties had to be separated by liquid chromatography prior to mass

50 spectrometric detection. The effect of different modifiers (i.e. acetic acid, ammonium acetate, ammoniun formate and formic acid), eluent pH and LC columns with different stationary phases were tested for the best chromatographic separation of the analytes. The pH of the eluents also affects to the ionisation efficiency of each analyte and thus the overall sensitivity of LC/MS methods. Here, we had to compromise when choosing the best LC/MS conditions for the cocktail assays. In study III, a mixture of 10 mM ammonium acetate and 1% formic acid was considered to be the best aqueous phase, especially for the strongly tailing desethylamodiaquine, although the best general sensitivity was obtained when the 0.1% acetic acid was used as the aqueous phase and methanol as the organic phase eluent. The use of methanol or acetonitrile may have a different effect to the ionization efficiency, and in some cases the selectivity of the column and thus separation efficiency as well may be improved by changing the acetonitrile to methanol. However, the use of acetonitrile leads to decreased column back pressures in comparison to methanol gradients. In study IV, the very poor ionization efficiency of repaglinide and losartan and their metabolites in basic conditions and the weak retention of nicotine and its metabolite cotinine in acidic conditions were the starting points for the development of LC method for the analytes. We sought to improve the retention of cotinine by using buffer solutions such as ammonium formate or ammonium acetate, but these ion pairing reagents decreased the detection sensitivity of the losartan and repaglinide metabolites. The best LC/MS performance was achieved with 0.1% acetic acid as the aqueous phase and using acetonitrile as organic phase eluent instead of methanol. The Waters BEH Shield RP18 (and the corresponding XBridge Shield RP18 with the HPLC system) column was substantially better than the other columns tested. In the case of in vivo cocktail, weak retention of cotinine required a low organic solvent content at the beginning of the gradient, whereas a high proportion of organic eluent was applied for the effective elution of losartan and repaglinide metabolites. In both studies the slope of the gradient was mainly determined by the separation of three isobaric metabolites of omeprazole. Human in vitro omeprazole hydroxylations to the 5- and 3- positions and sulfone metabolites responded with the same LC/MS/MS detection reaction optimized to hydroxymetabolites, so an adequate chromatographic separation step is needed to detect of all three metabolites. As the formation of 5-hydroxylation is activated by CYP2C19, whereas the 3-hydroxylation and sulfonation are formed via CYP3A4, it is of very high importance that these three metabolites are chromatographically well separated before the MS/MS-detection. The need for this separation may increase

51 Fig 10. The structures of the in vivo cocktail drugs and their CYP-speciﬁc metabolites.

52 the analysis time per sample in the assays utilizing omeprazole metabolites as CYP activity probes; this was the case when using HPLC in urine sample analysis in study IV. In a recent assay [134] the 5-hydroxyomeprazole was used as a probe metabolite for CYP2C19 from human liver microsomes with a very fast chromatographic method, and only one LC/MS/MS peak was detected with the MRM reaction from m/z 362 to m/z 214, suggesting that the responses of all three metabolites were mixed into a single peak, leading to the summarized response from activities of CYP3A4 and CYP2C19. In the N-in-one assays, the differences in concentrations between various CYP specific substrates and their probe metabolites may be as high as hundred fold and thus the need for high linear detection response is vital. However, the CYP interaction assays are typically run using only relative chromatographic responses for the same probe metabolites (or responses between the analyte and internal standard) from run to run, without testing the linear range of the method’s response by creating calibration curves with spiked standard samples. The method must also be capable for giving a specific response for each analyte and the result obtained must specifically reflect the activity of each CYP isoform, with no interference from the background or the probe metabolites present for other enzymes. Therefore, two different test approaches were utilized to obtain the effect of matrix components on the ionization response of the urine analytes, whereas only post-column infusion method was used for the in vitro probe metabolites. Another method employed with the urine matrix was the comparison of the LC/MS/MS peak intensities obtained from the samples spiked to same analyte concentration for urine and blank water. The observed enhancement or suppression of the ionization was between 6% and 25% with the urine matrix, and with the in vitro cocktail matrix the decrease in the detected baseline was no more than 10%. The results suggest the matrix effect is not significant in either case. The dilution of the urine samples was considered sufficient to keep the matrix effect at an acceptable level, whereas in the case of in vitro cocktail samples an even higher matrix effect would have been acceptable, as the method was used for comparing the peak areas of the samples with a similar matrix. Knowing the limits of the analytical method performance is highly important for estimating the reliability of the obtained results. In most of the in vitro N-in-one CYP screening assays the analytical methods used are not adequately validated and, even more alarmingly, some of the publications describing the assays give no information whatsoever regarding analytical method performance, decreasing the credibility of the functionality of the assays. In our studies, the analytical methods developed for cocktail assays were validated for linear range, detection limit, accuracy and precision for each

53 Table 5. Performance of the LC/MS/MS assay A, in vitro cocktail [III].

Compound* LoD (nM) Range (nM) Accuracy (%) Precision (%) Desethylamodiaquine 10 10-2000 (R2 = 0.992) 90-106 4-11 Hydroxymelatonin 2 2-4000 (R2 = 0.992) 93-107 3-15 Dextrorphan 1 1-1000 (R2 = 0.985) 85-110 1-16 Hydroxycoumarin 2 2-4000 (R2 = 0.998) 89-108 1-9 Hydroxybupropion 0.4 1-2000 (R2 = 0.993) 87-115 2-15 Hydroxychlorzoxazone 15 30-6000 (R2 = 0.994) 88-105 2-15 5-Hydroxyomeprazole 2 2-2000 (R2 = 0.992) 94-109 2-10 Hydroxytolbutamide 0.6 1.5-3000 (R2 = 0.999) 93-113 1-13 1-Hydroxymidazolam 0.2 1-1000 (R2 = 0.999) 93-108 1-14 6β-hydroxytestosterone 6 15-1500 (R2 = 0.985) 98-116 4-14 Omeprazole sulfone 0.4 2-1000 (R2 = 0.990) 88-107 4-11 * DeM-OME and 3-OH-OME were not available as standards. metabolite. Precision and accuracy were calculated at 7-11 and 6-8 concentrations within the speciﬁed ranges, in III and IV, respectively. The amount of parallel samples was three (n=3) in both of the studies. In the case of the in vitro cocktail assay [III], the validation was made with two different LC/MS/MS systems from different manufacturers, and showed a good applicability of the method. The performance of the LC/MS/MS method in the assay A (Waters instruments) is presented in Table 5 and the performance of assay B (Thermo Finnigan instrumentation) is presented in Table 3 in III. Fourteen compounds were analyzed from a single LC/MS/MS run. However, the Finnigan TSQ Quantum Discovery MAX triple quadrupole instrument was incapable of switching polarities, and thus the in vitro cocktail samples in assay B were analyzed with two different runs, as the negative ion mode had to be used for analyzing the hydroxychlorzoxazone alone. The detection limits were between 0.2 and 30 nM and the linear ranges obtained were 3 or 4 orders of magnitude with both of the instrumentations used. The accuracies were 85-116% and obtained standard deviations for precision were below 16%, thus at an acceptable level for screening-type analysis. The results obtained with both of the tested instrumentations were somewhat similar and no clear difference in the general detection sensitivity was obtained. However, the sensitivity for the metabolites with the lowest abundance was better with the Waters instrumentation and the capability of polarity switching made the Waters instrumentation (assay A) a better choice for the analysis described in study III. In study IV, twelve CYP-speciﬁc probe metabolites and their nine parent drugs from human urine were analyzed. The performance of the LC/MS/MS method utilizing UPLC is presented in Table 6. The HPLC method (assay B) was developed only for comparison

54 purposes, and thus only the detection limits for the analytes were estimated with the assay. The detection limits obtained with the UPLC/MS/MS method (assay A) were generally 0.4 ng/ml or less in urine, the only exception being hydroxychlorzoxazone and its substrate chlorzoxazone, for which the detection limits were 2.0 and 1.0 ng/ml, respectively. The highest detection sensitivity was obtained for 1-hydroxymidazolam, only 0.05 ng/ml in urine. The detection limits with the traditional HPLC method were 2 - 5 times higher than with the method utilizing UPLC, the LoDs being between 0.4 and 4.0 ng/ml. When taking into account the variables affecting non-chromatography-based detection sensitivity, such as injection volume used and the difference in sensitivities of the mass spectrometers used, about 1.5 - 3 times higher sensitivities were obtained using UPLC instead of HPLC. When comparing the performances of the UPLC and HPLC systems in the in vivo cocktail assay (Figure 11), superior chromatographic peak shapes were obtained when UPLC was utilized, leading to clearly increased chromatographic separation and higher detection sensitivity. The most signiﬁcant change in peak shape was obtained for losartan metabolite E3174, while the increase in detection sensitivity was most striking for omeprazole sulfone. The signal-to-noise ratio was over 100 times better compared to HPLC, which is probably due to the decreased matrix effect. In addition, the analysis speed was three times faster with UPLC in comparison to the corresponding method based on traditional HPLC. When increasing the speed of chromatographic run, the data acquisition speed (data points/s) of the detector must be taken into account, so that at least 8-10 time points per chromatographic peak should be collected to obtain the adequate shape and correct intensity (peak area) for the chromatographic peak. When performing exact quantitative work, the data acquisition rate requirement is normally stated as 12-15 data points per peak. When analyzing data concerning, for example, ten MRM-reactions at the same time, with dwell time and inter-scan delay being 100 ms and 20 ms, respectively, this would mean one data point for each reaction every 1.2 s, and if 10 data points over a chromatographic peak is desired, the rate is not fast enough unless the peak width is 12 s or more. In modern ultra fast chromatography, the peak widths may well be 2-3 s, thus requiring a clearly faster data acquisition rate than that calculated above. However, some modern triple quadrupole or TOF mass spectrometers can easily collect data points in times such as 10 ms per compound. Therefore, data acquisition must be optimized so that the number of data points / chromatographic peaks meets the criteria set above. If not, this can be improved by increasing the chromatographic run time to

55 Table 6. Performance of the LC/MS/MS assays [IV] (for assay B (HPLC) only the detection limit was estimated). Precision and accuracy are calculated at 6-8 concentrations (n=3) within the speciﬁed range.

Compound Assay A Assay B Retention time Range Accuracy Precision LoD Retention time LoD (min) (ng/ml) (%) (std. dev %) (ng/ml) (min) (ng/ml) Melatonin 2.03 1-1000 96-103 1-9 0.4 5.84 0.5 Hydroxymelatonin 1.58 1-2000 86-115 1-8 0.2 4.45 0.5 Nicotine 0.62 0.4-1000 86-113 4-14 0.2 0.83 2.0 Cotinine 0.78 0.4-2000 99-105 2-13 0.1 0.93 2.0 Bupropion 1.94 0.4-1000 88-104 1-11 0.2 4.20 1.0 Hydroxybupropion 1.67 1-1000 90-115 1-14 0.4 3.39 1.0 Repaglinide 3.88 0.4-1000 84-113 4-15 0.1 10.30 0.4 Hydroxyrepaglinide 3.01 a) a) a) a) 9.62 a) Losartan 2.87 0.4-1000 87-117 2-15 0.1 9.48 0.5 E3174 3.07 0.4-1000 92-107 1-7 0.2 10.15 1.0 Omeprazole 2.10 0.4-1000 90-103 1-11 0.1 6.15 0.4 Demethylomeprazole 1.38 a) a) a) a) 3.66 a) 5-hydroxyomeprazole 1.76 0.4-2000 87-103 1-12 0.2 5.35 0.5 3-hydroxyomeprazole 1.98 a) a) a) a) 6.05 a) Omeprazole sulphone 2.51 0.4-2000 95-112 1-8 0.2 6.25 0.8 Dextromethorphan 2.17 0.4-1000 88-101 3-9 0.2 5.41 0.4 Dextrorphan 1.58 1-1000 90-114 1-9 0.2 3.91 0.4 Chlorozoxazone 2.51 4-1000 91-118 1-12 1.0 7.30 2.0 6-hydroxychlorzoxazone 1.66 4-2000 80-117 1-15 2.0 4.74 4.0 Midazolam 2.37 1-1000 82-107 6-13 0.4 5.86 1.0 1-hydroxymidazolam 2.34 0.2-1000 90-115 1-8 0.05 6.89 0.2 a) not available as standard 56 Fig 11. SRM chromatograms for the analyzed metabolites with a method based on a) HPLC and b) UPLC. The data is acquired from urine sample collected from one study person dosed with nine drugs simultaneously. 3-Hydroxyomeprazole is not included as it was not present in the samples, and not available as a standard [IV]. obtain wider peaks, reducing the scan time (dwell time) used for one data point, or decreasing the number of overlapping MRM-detection reactions. The requirements for fast data acquisition are even stricter when analyzing both positive and negative ions at the same time. The reason for using this so-called polarity switching in our N-in-one cocktail assays was the analysis of CYP2E1 probe metabolite hydroxychlorzoxazone, which is not ionized in positive ion mode ESI but required analysis in negative ion mode [116, 118, 134, 136, 137]. Although, switching the voltages between opposite polarities led to additional delays between the scans, the amount of data points collected per

57 chromatographic peak was increased by decreasing the simultaneous MRM-reactions by using separate detection time windows for each analyte. This enabled detection of 11-14 analytes from each run in a short analysis time. The UPLC/MS/MS method developed for the analysis of twelve common probe metabolites for the nine most important drug metabolizing CYP isoforms enables completely non-invasive sample collection from the patients since the urine sample is sufﬁcient for the analysis. The obtained detection limits were very low, suggesting the possibility of reducing the N-in-one administered dose levels even lower than those used here. The LC/MS/MS method developed for the in vitro metabolic screening assay has proven to be a very robust and reliable method and has been used in a large number of CYP-inhibition screening studies with many different matrices.

Solid phase extraction tests for urine N-in-one assay

Different solid phase extraction methods were tested (unpublished data) in the process of developing a urinary based in vivo cytochrome P450 interaction / phenotyping cocktail. Figure 12 presents the recoveries obtained by combining the two elutions of the ion exchange cartridges together with the only elution from HLB. All ten probe metabolites (two probes for CYP3A4 and one for each of the remaining primary CYP isoenzymes) were detected solely from the HLB extract when only the last elution was collected. The recoveries from the HLB extract were 49 - 62%, with the exception of cotinine, hydroxymelatonin and hydroxybupropion, whose recoveries were 16%, 34% and 41%, respectively. Adjustment of elution solvent pH might have improved recoveries of some compounds, but probably also decreased recoveries for other compounds. When the cation exchange cartridge MCX was used and only the last elution was collected, the recoveries were 36 - 67%. The exceptions were cotinine, hydroxychlorzoxazone and hydroxymelatonin. Recoveries for cotinine and hydroxychlorzoxazone were 14% and 1%, respectively, whereas hydroxymelatonin was not detected. When two elution steps (wash step 2 using methanol and pH-adjusted ﬁnal elution) were combined (Figure 12), the recoveries for the hydroxylations of chlorzoxazone and melatonin increased substantially, whereas for E3174 (losartan acid) the recovery increased less, and for the rest of the compounds the recoveries remained about the same. With WCX, the weak cation exchange cartridge, dextrorphan was the only metabolite with a decent recovery (55%), as it was the most alkaline compound. Most of the compounds were not even detected and only 1% recoveries were obtained

58 for the detected compounds. However, when the two elutions were combined, the recoveries were 21 - 94%, except for hydroxychlorzoxazone and hydroxymelatonin, whose recoveries were 6% and 5%, respectively. As expected, the best recoveries were obtained for the acidic compounds with the anion exchange cartridge, MAX. When only the last elution was collected, the recoveries were 52 - 124%, except for cotinine, dextrorphan and the hydroxylations of melatonin, midazolam and bupropion, for which the recoveries were below 3%. Again, the same trend of quite equal recoveries within the compounds was obtained when the two elutions were combined. E3174 (losartan acid) seemed to be the only compound with strong acidic properties since it had the best recovery (71%) with WAX, when only the last elution was collected. Demethylomeprazole had 14% recovery, whereas the recoveries for the rest of the detected compounds were below 5%. When the elutions were combined, the recoveries were 42 - 76%, except for cotinine, dextrorphan and hydroxybupropion, whose recoveries were 4 - 21%. In general, the function of SPE cartridges was as expected. Recoveries of the basic compounds were higher when the cation exchange cartridges were used, and for the acidic compounds, the anion exchange cartridges were better. Some of the compounds with both acidic and basic properties, for example, omeprazole sulfone, E3174 and hydroxyrepaglinide, had recoveries of 50 - 80% in both cation (MCX) and anion (MAX) exchange cartridges. In general, the best SPE sorbent for the ten different kinds of metabolites was the universal sorbent HLB, the extraction procedure being simple and fast. The ion exchange cartridges also proved to be usable for most of the compounds when both of the elutions were collected for analysis. However, overall sensitivity with SPE was somewhat poor, the recoveries of the probe metabolites with HLB being only 16 - 62%, and thus suggesting the use of urine as such or diluted with ultra-pure water when analysing the N-in-one in vivo cocktail samples.

Single analyte assay [V]

In the last part of the study, a UPLC/MS/MS method was developed for analysis of an antidiabetic drug repaglinide from the samples collected to study its perfusion across placenta membranes. Repaglinide was one of the drugs used in the N-in-one in vivo cocktail study [IV] and thus the chromatography and mass spectrometric conditions for repaglinide were rather well-known from the earlier study and only some adjustments

59 60 elutions. combined two the from obtained recoveries SPE The 12. Fig

150

HLB MCX WCX MAX WAX 124 125

100 94

79 76 73 71 75 69 68 67 65 66 62 64 6264 62 58 59 59 60 58 Recovery, % Recovery, 55 5656 57 57 53 53 52 49 49 50 50 46 42 43 4139 34 36 31

25 21 21 16 17 14 13 6 4 5 3 0

E3174 (2C9) Cotinine (2A6) OH-CLZ (2E1) OH-Bup (2B6) OH-Mela (1A2) OH-MDZ (3A4) SO2-Ome (3A4) OH-Repa (2C8) O-dem-DXM (2D6) O-dem-Ome (2C19) were needed for obtaining a competent LC/MS performance. Antipyrine was used as a reference drug for passive diffusion-dependent placental perfusion, and due to high concentration a UPLC/PDA method was developed to analyze antipyrine. In contrast, a very high detection sensitivity was required for repaglinide, and thus solid phase extraction was used as a sample preparation step before the UPLC/MS/MS analysis. The analytical methods were validated for the linear range, detection limit, accuracy and precision. During method development, warfarin was chosen as an internal standard from the group of several tested compounds because of its similar behavior in solid phase extraction with repaglinide. A Waters BEH Shield RP18 column was chosen for the method and the eluents used were 0.1% acetic acid and acetonitrile. Ammonium formate (2 mM) was also tested as an aqueous phase, but better sensitivity was obtained for both compounds with acetic acid. In addition, the retention time difference increased using ammonium formate as the retention of IS decreased, whereas the retention of repaglinide increased. A linear gradient elution with 5 - 80% acetonitrile in 0 - 2 min was applied, followed by column equilibration with initial conditions for 1 minute. Different flow rates were tested for the optimum performance, the 0.7 ml/min flow rate being most suitable for the instrumentation used. When using UPLC columns packed with small size particles (1.7 µm), the increase in the mobile phase flow rate does not have a negative effect on the efficiency, as is the case with the HPLC columns packed with larger diameter particles (e.g. 5 µm), and thus the higher flow rates are encouraged to use with UPLC. However, increasing the mobile phase flow rate may have a negative influence on the detection sensitivity of ESI mass spectrometry, due to non-effective mobile phase evaporation. Selecting the right SPE sorbent depends strongly on the physicochemical properties of the analytes, and is thus the most important factor affecting the SPE performance. In addition, the sample matrix and the interactions with both the sorbent and analyte affect SPE performance [3]. The sample matrix used in the method development phase and for the preparation of the standard and quality control (QC) samples should be exactly the same or as close as possible to the analysis samples. However, this is not always possible, and some compromises have to be made when aiming at the closest alternative. Here, the matrix in the samples was a Krebs-Ringer solution containing BSA (bovine serum albumine), which was prepared in-house for the method development and STD/QC samples, and thus not exactly the same as in the real samples. The UV-chromatograms of the unknown samples were very similar to those of the standards, suggesting the

61 in-house prepared matrix solution was very close to the matrix of the real samples. Because of the assumption that the physicochemical properties of the substrate, repaglinide, were somewhat similar than to its hydroxymetabolite, the two most suitable SPE sorbents for OH-repaglinide were chosen from the earlier in vivo tests (Figure 12). Thus, Oasis HLB and MAX cartridges were tested with different wash and elution steps utilizing different combinations of organic solvents such as methanol, acetonitrile, isopropanol, ethyl acetate and acetone. The MAX anion-exchange sorbent was observed to be better than the universal sorbent HLB for both of the compounds, repaglinide and internal standard warfarin. The recovery of the MAX solid phase extraction method was 88 - 115% over the QC-range (50 - 5000 pg/ml, n=3 at each concentration) and for the warfarin (IS), the recovery was 105% (n=9). The selectivity of the UPLC/MS/MS method was confirmed with the matrix effect tests by means of post-column infusion method at 1 and 10 ng/ml repaglinide. The performance of the developed UPLC/MS/MS method is presented in Table 7. The linear range was 2 - 50 000 pg/ml and the accuracies at each standard concentration were 90 - 113%. The precisions at each level were 3 - 12% and the accuracies obtained from each individual QC samples were 90 - 112%. The lower limit of quantitation (2 pg/ml) was the same as the limit of detection. The developed and qualified UPLC/MS/MS method was a very sensitive and specific method for repaglinide analysis from the placenta perfusion samples. Despite of the competent results obtained with placenta samples, the suitability of the sample preparation method must be tested before utilizing the SPE method for other sample matrices. The UPLC/PDA method developed for antipyrine analysis was also very specific and sensitive: the limit of detection was 0.2 µg/ml and the obtained accuracies (n=3) and precisions (n=3) at each standard concentration were 92 - 113% and 1 - 10%, respectively, over the linear range 0.2 - 500 µg/ml.

62 Table 7. Performance of the single assay analysis UPLC/MS/MS method [V]. The accuracy and precision were calculated at each standard concentration (n=3). The calibration curve was created using linear ﬁtting and 1/x weighting, the obtained correlation coefﬁcient being > 0.997.

Concentration Accuracy Precision (pg/ml) (%) (%) 2 98.3 2.8 5 93.2 3.5 10 103.6 7.5 20 104.4 11.5 50 109.6 6.6 100 113.4 3.5 200 92.4 5.8 500 102.0 2.9 1000 94.2 4.7 2000 90.2 5.3 5000 98.1 9.3 10000 101.3 8.9 20000 98.4 6.6 50000 101.1 5.5

5.2.2 Comparison of N-in-one assay and single analyte analysis

When comparing the multi-analyte analysis i.e. N-in-one assay [IV] utilizing UPLC and the single analyte analysis (SAA) [V], the most striking difference was the ease of method development phase for the SAA, as the sample preparation step, instrument parameters and chromatography needed to be optimized for only one analyte (in addition to the internal standard). Furthermore, the analyte, repaglinide, was used in both of the assays, and thus made the method development even easier since the physicochemical properties of the analyte and the LC/MS conditions were already somewhat familiar. The chromatography for repaglinide and warfarin (IS) in the single analyte analysis

63 was adjusted without the boundaries created by the other analytes present in the cocktail sample. The weak retention of cotinine, the separation of three different omeprazole metabolites and the tailing of losartan were avoided, and thus a relatively fast UPLC method was developed for repaglinide. Similarly, the MS parameters, such as capillary voltage and desolvation temperature, were tuned only for the repaglinide (and warfarin) without compromises made in the N-in-one cocktail assay, thus increasing the detection sensitivity. The decreased number of simultaneous MRM reactions and, more precisely, the lack of polarity switching, also increased the sensitivity of the MS detection. All in all, the analytic method developed for the single analyte assay was faster and much more sensitive than the method used in the N-in-one assay. The retention time of repaglinide was 3.9 min in the N-in-one assay, whereas the retention time in the SAA was 1.7 min; thus the SAA method was more than twice as fast. The limit of detection for repaglinide in the SAA was 50 times lower than in the N-in-one assay, the LoDs being 2 pg/ml and 100 pg/ml, respectively. The much lower limit of detection in the SAA was due to the different sample preparation steps used, in addition to the better LC/MS optimization.

64 6 Conclusions

LC/MS was used for a qualitative and quantitative analysis of plant and drug metabolism products. Various LC/MS instrumentations were used in a versatile fashion and different LC/MS/MS systems were compared to obtain the best possible analytical performance. Of the different chromatographic systems, the UPLC was confirmed to be a better choice for simultaneous analysis of several compounds from a cocktail, in comparison to traditional HPLC. Triple quadrupole mass spectrometers were used for quantitative work and for more detailed structure elucidation, whereas the TOF instrument proved to be a powerful tool for screening purposes. However, the ion trap and Q-TOF instruments are able to provide detailed MS/MS data and similarly specific MRM-detection, and thus could be utilized for detailed identification and in analysis of the N-in-one assay, although the ion trap instruments are not capable for fast data acquisition, and thus unsuitable for use with UPLC. In addition, modern time-of-flight mass spectrometers could be used for quantitative work due to the introduction of dynamic range enhancement systems and improved ADC/TDC technology. As a result of this study, new qualified analytical methods were obtained for the new and more comprehensive in vitro and in vivo interaction cocktail assays. With these developed analytical methods, the studying of drugs and drug candidates affecting to CYP-enzyme activities is remarkably faster in comparison to older methods used. More extensive information about the factors affecting the whole CYP-enzyme family is also obtained simultaneously from a single assay. In this study, several new in vivo metabolites were also identified for the CYP2B6-selective probe drug, bupropion, after human administration. In addition, a specific and very sensitive analytical method was developed for the analysis of the CYP2C8 activity model drug, repaglinide, in the the human in vitro placental perfusion samples from the study model describing the transfer of compounds between the maternal and fetal blood flow through the placenta during pregnancy. In the part of the study concentrating on flavonoid secondary metabolism, the collision-induced radical cleavage of flavonoid glycosides was observed to be a suitable tool for structure elucidation of glycosylated flavonols. However, there are a few factors, such as instrument parameters, to be considered when analyzing the obtained MS/MS data. Most of the samples used in the study were complex mixtures and thus the utilization of different sample preparation methods such as SPE and PP prior to LC/MS

65 analysis had a notable implication for the obtained results. Some additional studies could be performed to provide a deeper look in the observed results or to improve the developed methods. For example, comparison of the bupropion metabolite proﬁle from plasma or serum with the observed urinary proﬁle would be interesting. The throughput of the in vitro interaction cocktail analysis could be improved further, even to below one minute per run. This would be especially facilitated by changing the CYP2C19 substrate omeprazole used, which forms three metabolites requiring clear chromatographic separation, to some other suitable substrate such as mephenytoin, producing only one target metabolite (4-hydroxymephenytoin). Moreover, the performance of the in vivo interaction cocktail could perhaps be further improved through sample preparation utilizing different SPE treatments (both cation and anion exchange) for the same sample, even though clearly increasing time required for sample preparation.

66 References

1. Cuyckens F & Claeys M (2004) Mass spectrometry in the structural analysis of flavonoids. J Mass Spectrom 39(1): 1–15. 2. Korfmacher W (2005) Principles and applications of LC-MS in new drug discovery. Drug Discov Today 10(20). 3. Nováková L & Vlckováˆ H (2009) A review of current trends and advances in modern bio-analytical methods: Chromatography and sample preparation. Anal Chim Acta 656: 8–35. 4. Guillarme D, Schappler J, Rudaz S & Veuthey JL (2010) Coupling ultra-high-pressure liquid chromatography with mass spectrometry. Trends Anal Chem 29(1): 15–27. 5. Coleman M (2005) Human Drug Metabolism: An Introduction. John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England. 6. Pelkonen O, Turpeinen M, Hakkola J, Honkakoski P, Hukkanen J & Raunio H (2008) Inhibition and induction of human cytochrome P450 enzymes: current status. Arch Toxicol 82: 667–715. 7. Walsky R & Obach R (2004) Validated assays for human cytochrome P450 activities. Drug Metab Dispos 32(6): 647–660. 8. Dierks E, Stams K, Lim HK, Cornelius G, Zhang H & Ball S (2001) A method for the simultanous evaluation of the activities of seven major human drug-metabolizing cytochrome P450s using an in vitro cocktail of probe substrates and fast gradient liquid chromatography tandem mass spectrometry. Drug Metab Dispos 29: 23–29. 9. Testino S & Patonay G (2003) High-throughput inhibition screening of major human cytochrome P450 enzymes using an in vitro cocktail and liquid chromatography-tandem mass spectrometry. J Pharm Biomed Anal 30: 1459–1467. 10. Alden P, Plumb R, Jones M, Rainville P & Shave D (2010) A rapid ultra-performance liquid chromatography/ tandem mass spectrometric methodology for the in vitro analysis of pooled and cocktail cytochrome P450 assays. Rapid Commun Mass Spectrom 24: 147–154. 11. Tanaka E, Kurata N & Yasuhara H (2003) How useful is the ‘cocktail approach’ for evaluating human hepatic drug metabolizing capacity using cytochrome P450 phenotyping probes in vivo? J Clin Pharmacol Ther 28: 157–165. 12. Fuhr U, Jetter A & Kirchheiner J (2007) Appropriate phenotyping procedures for drug metabolizing enzymes and transporters in humans and their simultaneous use in the “cocktail” approach. Clin Pharmacol Ther 81(2): 270–283. 13. Ghassabian S, Chetty M, Tattam B, Glen J, Rahme J, Stankovic Z, Ramzan I, Murray M & McLachlan A (2009) A high-throughput assay using liquid chromatography–tandem mass spectrometry for simultaneous in vivo phenotyping of 5 major cytochrome P450 enzymes in patients. Ther Drug Monit 31: 239–246. 14. Winkel-Shirley B (2002) Biosynthesis of flavonoids and effects of stress. Curr Opin Plant Biol 5: 218–223. 15. Koes R, Verweij W & Quattrocchio F (2005) Flavonoids: a colorful model for the regulation and evolution of biochemical pathways. Trends Plant Sci 10(5): 236–242. 16. Hertog M & Feskens E (1993) Dietary antioxidant flavonoids and risk of coronary heart disease: The Zutphen elderly study. Lancet 342(8878): 1007–1011.

67 17. Mojzisova G & Kuchta M (2001) Dietary flavonoids and risk of coronary heart disease. Physiol Res 50(6): 529–535. 18. Bazzano L, He J, Ogden L, Loria C, Vupputuri S, Myers L & Whelton P (2002) Fruit and vegetable intake and risk of cardiovascular disease in US adults: the first national health and nutrition examination survey epidemiologic follow-up study1–3. Am J Clin Nutr 76: 93–99. 19. Le Marchand L (2002) Cancer preventive effects of flavonoids—a review. Biomed Pharmacother 56: 296–301. 20. Birt D, Hendrich S & Wang W (2001) Dietary agents in cancer prevention: flavonoids and isoflavonoids. Pharmacol Ther 90: 157–177. 21. Pietta PG (2000) Flavonoids as antioxidants. J Nat Prod 63(7): 1035–1042. 22. Havsteen B (2002) The biochemistry and medical significance of the flavonoids. Pharmacol Ther 96: 67–202. 23. Nijveldt R, van Nood E, van Hoorn D, Boelens P, van Norren K & van Leeuwen P (2001) Flavonoids: a review of probable mechanisms of action and potential applications. Am J Clin Nutr 74: 418–425. 24. Song Y, Manson J, Buring J, Sesso H & Liu S (2005) Associations of dietary flavonoids with risk of type 2 diabetes, and markers of insulin resistance and systemic inflammation in women: A prospective study and cross-sectional analysis. J Am Coll Nutr 24(5): 376–384. 25. Varma S, Mizuno A & Kinoshita J (1977) Diabetic cataracts and flavonoids. Science 195(4274): 205–206. 26. Crozier A, Jaganath I & Clifford M (2009) Dietary phenolics: chemistry, bioavailability and effects on health. Nat Prod Rep 26(8): 1001–1043. 27. Seigler D (1998) Plant secondary metabolism, chapter 11, pp. 151–192. Kluwer Academic Publishers, Norwell, Massachusetts. 28. Jaakola L (2003) Flavonoid biosynthesis in bilberry (vaccinium myrtillus l.). Ph.D. thesis, Department of Biology, University of Oulu. 29. Murray J & Hackett W (1991) Dihydroflavonol reductase activity in relation to differential anthocyanin accumulation in juvenile and mature phase hedera helix l. Plant Physiol 97(1): 343–351. 30. Williams C (2006) Flavonoids : chemistry, biochemistry, and applications, chapter 13, pp. 749–856. CRC Press / Taylor & Francis Group, 6000 Broken Sound Parkway NW, Suite 300 Boca Raton. 31. Stobiecki M (2000) Application of mass spectrometry for identification and structural studies of flavonoid glycosides. Phytochemistry 54: 237–256. 32. De Rijke E, Zappey H, Ariese F, Gooijer C & Brinkman U (2003) Liquid chromatography with atmospheric pressure chemical ionization and electrospray ionization mass spectrometry of flavonoids with triple-quadrupole and ion-trap instruments. J Chromatogr A 984: 45–58. 33. Cuyckens F, Ma YL, Pocsfalvi G & Claeys M (2000) Tandem mass spectral strategies for the structural characterization of flavonoid glycosides. Analusis 28(10): 888–895. 34. Cuyckens F, Rozenberg R, de Hoffmann E & Claeys M (2001) Structure characterization of flavonoid O-diglycosides by positive and negative nano-electrospray ionization ion trap mass spectrometry. J Mass Spectrom 36: 1203–1210. 35. Justesen U (2001) Collision-induced fragmentation of deprotonated methoxylated flavonoids, obtained by electrospray ionization mass spectrometry. J Mass Spectrom 36: 169–178. 36. March R & Miao XS (2004) A fragmentation study of kaempferol using electrospray quadrupole time-of-flight mass spectrometry at high mass resolution. Int J Mass Spectrom

68 231: 157–167. 37. March R, Miao XS, Metcalfe C, Stobiecki M & L M (2004) A fragmentation study of an isoflavone glycoside, genistein-7-O-glucoside, using electrospray quadrupole time-of-flight mass spectrometry at high mass resolution. Int J Mass Spectrom 232: 171–183. 38. Cuyckens F & Claeys M (2005) Determination of the glycosylation site in flavonoid mono- O-glycosides by collision-induced dissociation of electrospray-generated deprotonated and sodiated molecules. J Mass Spectrom 40: 364–372. 39. Vukics V & Guttman A (2010) Structural characterization of flavonoid glycosides by multi-stage mass spectrometry. Mass Spec Rev 29: 1–16. 40. Ma YL, Li Q, Van den Heuvel H & Claeys M (1997) Characterization of flavone and flavonol aglycones by collision-induced dissociation tandem mass spectrometry. Rapid Commun Mass Spectrom 11: 1357–1364. 41. Domon B & Costello C (1988) A systematic nomenclature for carbohydrate fragmentations in FAB-MS/MS spectra of glycoconjugates. Glycoconj J 5: 397–409. 42. Cuyckens F, AA S, Pieters L & Claeys M (2002) Direct stereochemical assignment of hexose and pentose residues in flavonoid O-glycosides by fast atom bombardment and electrospray ionization mass spectrometry. J Mass Spectrom 37: 1272–1279. 43. Fabre N, Rustan I, de Hoffmann E & Quetin-Leclercq J (2001) Determination of flavone, flavonol, and flavanone aglycones by negative ion liquid chromatography electrospray ion trap mass spectrometry. J Am Soc Mass Spectrom 12: 707–715. 44. Ma YL, Vedernikova I, Van den Heuvel H & Claeys M (2000) Internal glucose residue loss in protonated O-diglycosyl flavonoids upon low-energy collision-induced dissociation. J Am Soc Mass Spectrom 11: 136–144. 45. Ma YL, Cuyckens F, Van den Heuvel H & Claeys M (2001) Mass spectrometric methods for the characterisation and differentiation of isomeric O-diglycosyl flavonoids. Phytochem Anal 12: 159–165. 46. Ferreres F, Llorach R & Gil-Izquierdo A (2004) Characterization of the interglycosidic linkage in di-, tri-, tetra- and pentaglycosylated flavonoids and differentiation of positional isomers by liquid chromatography/electrospray ionization tandem mass spectrometry. J Mass Spectrom 39: 312–321. 47. Li Q & Claeys M (1994) Characterization and differentiation of diglycosyl flavonoids by positive ion fast atom bombardment and tandem mass spectrometry. Biol Mass Spectrom 23(7): 406–416. 48. Hvattum E & Ekeberg D (2003) Study of the collision-induced radical cleavage of flavonoid glycosides using negative electrospray ionization tandem quadrupole mass spectrometry. J Mass Spectrom 38: 43–49. 49. Ablajan K, Abliz Z, Shang XY, He JM, Zhang RP & Shi JG (2006) Structural characterization of flavonol 3,7-di-O-glycosides and determination of the glycosylation position by using negative ion electrospray ionization tandem mass spectrometry. J Mass Spectrom 41: 352–360. 50. Davis B & Brodbelt J (2008) An investigation of the homolytic saccharide cleavage of deprotonated flavonol 3-O-glycosides in a quadrupole ion trap mass spectrometer. J Mass Spectrom 43: 1045–1052. 51. Geng P, Sun J, Zhang R, He J & Abliz Z (2009) An investigation of the fragmentation differences of isomeric flavonol-O-glycosides under different collision-induced dissociation based mass spectrometry. Rapid Commun Mass Spectrom 23: 1519–1524.

69 52. Lu L, Song FR, Tsao R, Jin YR, Liu ZQ & Liu SY (2010) Studies on the homolytic and heterolytic cleavage of kaempferol and kaempferide glycosides using electrospray ionization tandem mass spectrometry. Rapid Commun Mass Spectrom 24: 169–172. 53. Fura A, Shu YZ, Zhu M, Hanson R, Roongta V & Humphreys W (2004) Discovering drugs through biological transformation: Role of pharmacologically active metabolites in drug discovery. J Med Chem 47(18): 4339–4351. 54. Smith D & Obach R (2009) Metabolites in safety testing (MIST): Considerations of mechanisms of toxicity with dose, abundance, and duration of treatment. Chem Res Toxicol 22: 267–279. 55. James L, Mayeux P & Hinson J (2003) Acetaminophen-induced hepatotoxicity. Drug Metab Dispos 31(12): 1499–1506. 56. Laine J, Auriola S, Pasanen M & Juvonen R (2009) Acetaminophen bioactivation by human cytochrome P450 enzymes and animal microsomes. Xenobiotica 39(1): 11–21. 57. Parkinson A & Ogilvie B (2007) Casarett and Doull’s Toxicology - The Basic Science of Poisons, chapter Biotransformation of xenobiotics, pp. 181–304. The McGraw-Hill Companies, Inc. 58. Yan Z & Caldwell G (2001) Metabolism proﬁling, and cytochrome P450 inhibition & induction in drug discovery. Curr Top Med Chem 1: 403–425. 59. Gibson GG & Skett P (2001) Introduction to drug metabolism, chapter Glucuronide conjugation reactions, pp. 64–71. Nelson Thornes publishers, 3rd edition. 60. Tukey R & Strassburg C (2001) Genetic multiplicity of the human UDP- glucuronosyltrans- ferases and regulation in the gastrointestinal tract. Mol Pharmacol 59(3): 405–414. 61. Patten C (2006) New technologies for assessing UDP-glucuronosyltransferase (UGT) metabolism in drug discovery and development. Drug Discov Today: Technologies 3(1): 73–78. 62. Williams J, Hurst S, Bauman J, Jones B, Hyland R, Gibbs J, Obach R & Ball S (2003) Reaction phenotyping in drug discovery: Moving forward with conﬁdence? Curr Drug Metab 4: 527–534. 63. Dorne J, Walton K & Renwick A (2005) Human variability in xenobiotic metabolism and pathway-related uncertainty factors for chemical risk assessment: a review. Food Chem Toxicol 43: 203–216. 64. Obach R (2000) Inhibition of human cytochrome P450 enzymes by constituents of St. John’s wort, an herbal preparation used in the treatment of depression. J Pharmacol Exp Ther 294(1): 88–95. 65. Markowitz J, Donovan J, DeVane L, Taylor R, Ruan Y, Wang JS & Chavin K (2003) Effect of St John’s wort on drug metabolism by induction of cytochrome P450 3A4 enzyme. JAMA 290(11): 1500–1504. 66. Rendic S & Di Carlo F (1997) Human cytochrome P450 enzymes: A status report summarizing their reactions, substrates, inducers and inhibitors. Drug Metab Rev 29(1 & 2): 413–580. 67. Zhou SF & Lai X (2008) An update on clinical drug interactions with the herbal antidepres- sant St. John’s wort. Curr Drug Metab 9: 394–409. 68. Guengerich F & Rendic S (2002) Update information on human P450s. Drug Metab Rev 34(1): 7–15. 69. Ioannides C & Lewis D (2004) Cytochromes P450 in the bioactivation of chemicals. Curr Top Med Chem 4(16): 1767–1788.

70 70. Danielson P (2002) The cytochrome P450 superfamily: Biochemistry, evolution and drug metabolism in humans. Curr Drug Metab 3(6): 561–597. 71. Lewis D (2003) Human cytochromes P450 associated with the phase 1 metabolism of drugs and other xenobiotics: A compilation of substrates and inhibitors of the CYP1, CYP2 and CYP3 families. Curr Med Chem 10: 1955–1972. 72. Nelson D, Koymans L, Kamataki T, Stegeman J, Feyereisen R, Waxman D, Waterman M, Gotoh O, Coon M, Estabrook R, Gunsalus I & Nebert D (1996) P450 superfamily: update on new sequences, gene mapping, accession numbers and nomenclature. Pharmacogenetics 6(1): 1–42. 73. Tolonen A, Turpeinen M & Pelkonen O (2009) Liquid chromatography-mass spectrometry in in vitro drug metabolite screening. Drug Discov Today 14(34): 120–133. 74. Kataoka H (2003) New trends in sample preparation for clinical and pharmaceutical analysis. Trends Anal Chem 22(4): 232–244. 75. Chen Y, Guo Z, Wang X & Qiu C (2008) Sample preparation. J Chromatogr A 1184: 191–219. 76. Fontanals N, Marcé R & Borrull F (2005) New hydrophilic materials for solid-phase extraction. Trends Anal Chem 24(5): 394–406. 77. Swartz M (2005) UPLC: An introduction and review. J Liq Chrom Relat Tech 28: 1253– 1263. 78. Churchwell M, Twaddle N, Meeker L & Doerge D (2005) Improving LC–MS sensitivity through increases in chromatographic performance: Comparisons of UPLC–ES/MS/MS to HPLC–ES/MS/MS. J Chromatogr B 825: 134–143. 79. O’Connor D & Mortishire-Smith R (2006) High-throughput bioanalysis with simultaneous acquisition of metabolic route data using ultra performance liquid chromatography coupled with time-of-flight mass spectrometry. Anal Bioanal Chem 385: 114–121. 80. Wu N & Clausen A (2007) Fundamental and practical aspects of ultrahigh presssure liquid chromatography for fast separations. J Sep Sci 30: 1167–1182. 81. Cech N & Enke C (2001) Practical implications of some recent studies in electrospray ionization fundamentals. Mass Spec Rev 20: 362–387. 82. Leinonen A, Kuuranne T & Kostiainen R (2002) Liquid chromatography/mass spectrometry in anabolic steroid analysis — optimization and comparison of three ionization techniques: electrospray ionization, atmospheric pressure chemical ionization and atmospheric pressure photoionization. J Mass Spectrom 37: 693–698. 83. Dams R, Huestis M, Lambert W & Murphy C (2003) Matrix effect in bio-analysis of illicit drugs with LC-MS/MS: Influence of ionization type, sample preparation, and biofluid. J Am Soc Mass Spectrom 14: 1290–1294. 84. Niessen W, Manini P & Andreoli R (2006) Matrix effects in quantitative pesticide analysis using liquid chromatography-mass spectrometry. Mass Spec Rev 25: 881–899. 85. Antignac JP, de Wasch K, Monteau F, De Brabander H, Andre F & Le Bizec B (2005) The ion suppression phenomenon in liquid chromatography–mass spectrometry and its consequences in the field of residue analysis. Anal Chim Acta 529: 129–136. 86. Kebarle P (2000) A brief overview of the present status of the mechanisms involved in electrospray mass spectrometry. J Mass Spectrom 35: 804–817. 87. Liang H, Foltz F, Meng M & Bennet P (2003) Ionization enhancement in atmospheric pressure chemical ionization and suppression in electrospray ionization between target drugs and stable-isotopelabeled internal standards in quantitative liquid chromatography/tandem

71 mass spectrometry. Rapid Commun Mass Spectrom 17: 2815–2821. 88. Tang K, Page J & Smith R (2004) Charge competition and the linear dynamic range of detection in electrospray ionization mass spectrometry. J Am Soc Mass Spectrom 15: 1416–1423. 89. Jessome L & Volmer D (2006) Ion suppression: A major concern in mass spectrometry. LCGC North America 24(5): 498–510. 90. Williamson L & Bartlett M (2007) Quantitative liquid chromatography/time-of-flight mass spectrometry. Biomed Chromatogr 21: 567–576. 91. Li X, Jiang G, Luo C, Xu F, Wang Y, Ding L & Ding CF (2009) Ion trap array mass analyzer: Structure and performance. Anal Chem 81: 4840–4846. 92. Mortishire-Smith R, Castro-Perez J, Yu K, Shockcor J, Goshawk J, Hartshorn M & Hill A (2009) Generic dealkylation: a tool for increasing the hit-rate of metabolite rationalization, and automatic customization of mass defect filters. Rapid Commun Mass Spectrom 23: 939–948. 93. Tiller P, Yu S, Castro-Perez J, Fillgrove K & Baillie T (2008) High-throughput, accurate mass liquid chromatography/ tandem mass spectrometry on a quadrupole time-of-flight system as a ‘first-line’ approach for metabolite identification studies. Rapid Commun Mass Spectrom 22: 1053–1061. 94. Leclercq L, Mortishire-Smith R, Huisman M, Cuyckens F, Hartshorn M & Hill A (2009) Isoscore: automated localization of biotransformations by mass spectrometry using product ion scoring of virtual regioisomers. Rapid Commun Mass Spectrom 23: 39–50. 95. Zhang H, Zhang D & Ray K (2003) A software filter to remove interference ions from drug metabolites in accurate mass liquid chromatography/mass spectrometric analyses. J Mass Spectrom 38: 1110–1112. 96. Zhang H, Zhang D, Ray K & Zhu M (2009) Mass defect filter technique and its applications to drug metabolite identification by high-resolution mass spectrometry. J Mass Spectrom 44: 999–1016. 97. Pelander A, Tyrkkö E & Ojanperä I (2009) In silico methods for predicting metabolism and mass fragmentation applied to quetiapine in liquid chromatography/time-of-flight mass spectrometry urine drug screening. Rapid Commun Mass Spectrom 23: 506–514. 98. Lasser K, Allen P, Woolhandler S, Himmelstein D, Wolfe S & Bor D (2002) Timing of new black box warnings and withdrawals for prescription medications. JAMA 287(17): 2215–2220. 99. Pelkonen O, Turpeinen M, Uusitalo J, Rautio A & Raunio H (2005) Prediction of drug metabolism and interactions on the basis of in vitro investigations. Basic Clin Pharmacol Toxicol 96: 167–175. 100. Frye R, Matzke G, Adedoyin A, Porter J & Branch R (1997) Validation of the five-drug ”Pittsburgh cocktail” approach for assessment of selective regulation of drug-metabolizing enzymes. Clin Pharmacol Ther 62: 365. 101. Streetman D, Bleakley J, Kim J, Nafziger A, Leeder J, Gaedigk A, Gotschall R, Kearns G & Bertino Jr J (2000) Combined phenotypic assessment of CYP1A2, CYP2C19, CYP2D6, CYP3A, N-acetyltransferase-2, and xanthine oxidase with the ”Cooperstown cocktail”. Clin Pharmacol Ther 68: 375–383. 102. Zhu B, Ou-Yang DS, Chen XP, Huang SL, Tan ZR, He N & Zhou HH (2001) Assessment of cytochrome P450 activity by a five-drug cocktail approach. Clin Pharmacol Ther 70: 455–461.

72 103. Chainuvati S, Nafziger A, Leeder J, Gaedigk A, Kearns G, Sellers E, Zhang Y, Kashuba A, Rowland E & Bertino J (2003) Combined phenotypic assessment of cytochrome P450 1A2, 2C9, 2C19, 2D6, and 3A, N-acetyltransferase-2, and xanthine oxidase activities with the ”Cooperstown 5+1 cocktail”. Clin Pharmacol Ther 74: 437–447. 104. Christensen M, Andersson K, Dalén P, Mirghani R, Muirhead G, Nordmark A, Tybring G, Wahlberg A, Yasar U & Bertilsson L (2003) The Karolinska cocktail for phenotyping of five human cytochrome P450 enzymes. Clin Pharmacol Ther 73: 517–528. 105. Jerdi M, Daali Y, Oestreicher M, Cherkaoui S & Dayer P (2004) A simplified analytical method for a phenotyping cocktail of major CYP450 biotransformation routes. J Pharm Biomed Anal 35: 1203–1212. 106. Sharma A, Pilote S, Bélanger P, Arsenault M & Hamelin B (2004) A convenient five-drug cocktail for the assessment of major drug metabolizing enzymes: a pilot study. Br J Clin Pharmacol 53: 288–297. 107. Tomalik-Scharte D, Jetter A, Kinzig-Schippers M, Skott A, Sörgel F, Klaassen T, Kasel D, Harlfinger S, Doroshyenko O, Frank D, Kirchheiner J, Bräter M, Richter K, Gramatté R & Fuhr U (2005) Effect of propiverine on cytochrome P450 enzymes: A cocktail interaction study in healthy volunteers. Drug Metab Dispos 33: 1859–1866. 108. Yin O, Lam S, Lo C & Chow M (2004) Rapid determination of five probe drugs and their metabolites in human plasma and urine by liquid chromatography/tandem mass spectrometry: application to cytochrome P450 phenotyping studies. Rapid Commun Mass Spectrom 18: 2921–2933. 109. Palmer J, Scott R, Gibson A, Dickins M & Pleasance S (2001) An interaction between the cytochrome P450 probe substrates chlorzoxazone (CYP2E1) and midazolam (CYP3A). Br J Clin Pharmacol 52: 555–561. 110. Zhang W, Han F, Guo P, Zhao H, Lin Z, Huang MQ, Bertelsen K & Weng N (2010) Simultaneous determination of tolbutamide, omeprazole, midazolam and dextromethorphan in human plasma by LC–MS/MS—a high throughput approach to evaluate drug–drug interactions. J Chromatogr B 878: 1169–1177. 111. Videau O, Delaforge M, Levi M, Thevenot E, Gal O, Becquemont L, Beaune P & Benech H (2010) Biochemical and analytical development of the CIME cocktail for drug fate assessment in humans. Rapid Commun Mass Spectrom 24: 2407–2419. 112. Youdim K & Saunders K (2010) A review of LC–MS techniques and high-throughput approaches used to investigate drug metabolism by cytochrome P450s. J Chromatogr B 878(16-17): 1326–1336. 113. Zhang T, Zhu Y & Gunaratna C (2002) Rapid and quantitative determination of metabolites from multiple cytochrome P450 probe substrates by gradient liquid chromatography– electrospray ionization-ion trap mass spectrometry. J Chromatogr B 780: 371–379. 114. Turpeinen M, Uusitalo J, Jalonen J & Pelkonen O (2005) Multiple P450 substrates in a single run: rapid and comprehensive in vitro interaction assay. Eur J Pharm Sci 24(1): 123–132. 115. Mohutsky M, Petullo D & Wrighton S (2005) The use of a substrate cassette strategy to improve the capacity and throughput of cytochrome P450 induction studies in human hepatocytes. Drug Metab Dispos 33(7): 920–923. 116. Lahoz A, Donato M, Picazo L, Gomez-Lechon M & Castell J (2007) Determination of major human cytochrome P450s activities in 96-well plates using liquid chromatography tandem mass spectrometry. Toxicol in Vitro 21: 1247–1252.

73 117. Dixit V, Hariparsad N, Desai P & Unadkat J (2007) In vitro LC-MS cocktail assays to simultaneously determine human cytochrome P450 activities. Biopharm Drug Dispos 28: 257–262. 118. Li X, Chen X, Li Q, Wang L & Zhong D (2007) Validated method for rapid inhibition screening of six cytochrome P450 enzymes by liquid chromatography–tandem mass spectrometry. J Chromatogr B 852: 128–137. 119. Rainville P, Wheaton J, Alden P & Plumb R (2008) Sub one minute inhibition assays for the major cytochrome P450 enzymes utilizing ultra-performance liquid chromatography/tandem mass spectrometry. Rapid Commun Mass Spectrom 22: 1345–1350. 120. Bonfiglio R, King R, Olah TV & Merkle K (1999) The effects of sample preparation methods on the variability of the electrospray ionization response for model drug compounds. Rapid Commun Mass Spectrom 13(12): 1175–1185. 121. Hsieh Y, Chintala M, Mei H, Agans J, Brisson JM, Ng K & Korfmacher W (2001) Quantitative screening and matrix effect studies of drug discovery compounds in monkey plasma using fast-gradient liquid chromatography/tandem mass spectrometry. Rapid Commun Mass Spectrom 15(24): 2481–2487. 122. Van Eeckhaut A, Lanckmans K, Sarre S, Smolders I & Michotte Y (2009) Validation of bioanalytical LC–MS/MS assays: Evaluation of matrix effects. J Chromatogr B 877: 2198–2207. 123. Zapesochnaya G & Kurkin V (1983) The flavonoids of the rhizomes of Rhodiola rosea. ii. a flavonolignan and glycosides of herbacetin. Chem Nat Compd 19(1): 21–29. 124. Kurkin V, Zapesochnaya G & Shchavlinskii A (1984) Flavonoids of the epigeal part of Rhodiola rosea. i. Chem Nat Compd 20(5): 623–624. 125. Kurkin V, Zapesochnaya G & Shchavlinskii A (1984) Flavonoids of the rhizomes of rhodiola rosea. iii. Chem Nat Compd 20(3): 367–368. 126. Zapesochnaya G, Kurkin V & Shchavlinskii A (1985) Flavonoids of the epigeal part of Rhodiola rosea. ii. structures of new glycosides of herbacetin and of gossypetin. Chem Nat Compd 21(4): 464–473. 127. Huck C, Buchmeiser M & Bonn G (2001) Fast analysis of flavonoids in plant extracts by liquid chromatography–ultraviolet absorbance detection on poly(carboxylicacid)-coated silica and electrospray ionization tandem mass spectrometric detection. J Chromatogr A 943: 33–38. 128. Stewart J, Berkel H, Parish R, Simar M, Syed A, Bocchini J, Wilson J & Manno J (2001) Single-dose pharmacokinetics of bupropion in adolescents: Effects of smoking status and gender. J Clin Pharmacol 41: 770–778. 129. Borges V, Yang E, Dunn J & Henion J (2004) High-throughput liquid chromatography– tandem mass spectrometry determination of bupropion and its metabolites in human, mouse and rat plasma using a monolithic column. J Chromatogr B 804: 277–287. 130. Bondarev M, Bondareva T, Young R & Glennon R (2003) Behavioral and biochemical investigations of bupropion metabolites. Eur J Pharmacol 474: 85–93. 131. Schroeder D (1983) Metabolism and kinetics of bupropion. J Clin Psychiatry 44: 79–81. 132. Taylor L, Singh R, Chang S & Johnson R (1995) The identification of in vitro metabolites of bupropion using ion trap mass spectrometry. Rapid Commun Mass Spectrom 9: 902–910. 133. Weaver R, Graham K, Beattie I & Riley R (2003) Cytochrome P450 inhibition using recombinant proteins and mass spectrometry/multiple reaction monitoring technology in a cassette incubation. Drug Metab Dispos 31(7): 955–966.

74 134. He F, Bi HC, Xie ZY, Zuo Z, Li JK, Li X, Zhao LZ, Chen X & Huang M (2007) Rapid determination of six metabolites from multiple cytochrome P450 probe substrates in human liver microsome by liquid chromatography/mass spectrometry: application to high-throughput inhibition screening of terpenoids. Rapid Commun Mass Spectrom 21: 635–643. 135. Ryu J, Song I, Sunwoo Y, Shon J, Liu K, Cha I & Shin J (2007) Development of the ”inje cocktail” for high-throughput evaluation of ﬁve human cytochrome P450 isoforms in vivo. Clin Pharmacol Ther 82(5): 531–540. 136. Bu HZ, Magis L, Knuth K & Teitelbaum P (2000) High-throughput cytochrome P450 (CYP) inhibition screening via cassette probe-dosing strategy. i. development of direct injection/online guard cartridge extraction/tandem mass spectrometry for the simultaneous detection of CYP probe substrates and their metabolites. Rapid Commun Mass Spectrom 14: 1619–1624. 137. Bu HZ, Magis L, Knuth K & Teitelbaum P (2001) High-throughput cytochrome P450 (CYP) inhibition screening via a cassette probe-dosing strategy. vi. simultaneous evaluation of inhibition potential of drugs on human hepatic isozymes CYP2A6, 3A4, 2C9, 2D6 and 2E1. Rapid Commun Mass Spectrom 15: 741–748.

75 76 Original articles

I Petsalo A, Jalonen J & Tolonen A (2006) Identification of flavonoids of Rhodiola rosea by liquid chromatography-tandem mass spectrometry. Journal of Chromatography A 1112: 224–231. II Petsalo A, Turpeinen M & Tolonen A (2007) Identification of bupropion urinary metabolites by liquid chromatography-mass spectrometry. Rapid Communications in Mass Spectrometry 21: 2547–2554. III Tolonen A, Petsalo A, Turpeinen M, Uusitalo J & Pelkonen O (2007) In vitro interaction cocktail assay for nine major cytochrome P450 enzymes with thirteen probe reactions and a single LC/MS/MS run; analytical validation and further testing with monoclonal P450 antibodies. Journal of Mass Spectrometry, 42: 960–966. IV Petsalo A, Turpeinen M, Pelkonen O & Tolonen A (2008) Analysis of nine drugs and their cytochrome P450 specific probe metabolites from urine by liquid chromatography-tandem mass spectrometry utilizing sub 2 µm particle size column. Journal of Chromatography A 1215: 107–115. V Tertti K, Petsalo A, Niemi M, Ekblad U, Tolonen A, Rönnemaa T, Turpeinen M, Heikkinen T & Laine K (2011) Transfer of repaglinide in the dually perfused human plancenta and the role of organic anion transporting polypeptides (OATPs). Manuscript.

Reprinted with permission of Elsevier (I, IV) & John Wiley and Sons (II, III). Original publications are not included in the electronic version of the dissertation.

77 78 ACTA UNIVERSITATIS OULUENSIS SERIES A SCIENTIAE RERUM NATURALIUM

558. Lappalainen, Niina (2010) The responses of ectohydric and endohydric mosses under ambient and enhanced ultraviolet radiation 559. Luojus, Satu (2010) From a momentary experience to a lasting one : the concept of and research on expanded user experience of mobile devices 560. Siirtola, Antti (2010) Algorithmic multiparameterised verification of safety properties : process algebraic approach 561. Lappi, Anna-Kaisa (2010) Mechanisms of protein disulphide isomerase catalyzed disulphide bond formation 562. Sarala, Marian (2010) Elongation of Scots pine seedlings under blue light depletion 563. Vance, Anthony (2010) Why do employees violate is security policies? : insights from multiple theoretical perspectives 564. Karppinen, Katja (2010) Biosynthesis of hypericins and hyperforins in Hypericum perforatum L. (St. John’s wort) – precursors and genes involved 565. Louhi, Pauliina (2010) Responses of brown trout and benthic invertebrates to catchment-scale disturbance and in-stream restoration measures in boreal river systems 566. Hekkala, Riitta (2011) The many facets of an inter-organisational information system project as perceived by the actors 567. Niittyvuopio, Anne (2011) Adaptation to northern conditions at flowering time genes in Arabidopsis lyrata and Arabidopsis thaliana 568. Leppälä, Johanna (2011) The genetic basis of incipient speciation in Arabidopsis lyrata 569. Kivelä, Sami, Mikael (2011) Evolution of insect life histories in relation to time constraints in seasonal environments : polymorphism and clinal variation 570. Kaartinen, Salla (2011) Space use and habitat selection of the wolf (Canis lupus) in human-altered environment in Finland 571. Hilli, Sari (2011) Carbon fractions and stocks in organic layers in boreal forest soils—impacts of climatic and nutritional conditions 572. Jokipii-Lukkari, Soile (2011) Endogenous haemoglobins and heterologous Vitreoscilla haemoglobin in hybrid aspen 573. Vuosku, Jaana (2011) A matter of life and death – Polyamine metabolism during zygotic embryogenesis pine

Book orders: Granum: Virtual book store http://granum.uta.fi/granum/ A 574 OULU 2011 A 574

UNIVERSITY OF OULU P.O.B. 7500 FI-90014 UNIVERSITY OF OULU FINLAND ACTA UNIVERSITATISUNIVERSITATIS OULUENSISOULUENSIS ACTA UNIVERSITATIS OULUENSIS ACTAACTA

SERIES EDITORS SCIENTIAESCIENTIAEA A RERUMRERUM Aleksanteri Petsalo NATURALIUMNATURALIUM

EDITOR IN CHIEF Professor Olli Vuolteenaho PUBLICATIONS EDITOR